Skip to main content

Full text of "Condensed Handbook Of Composition Input (book)"

See other formats





Condensed 

Handbook 

of 

Composition 

Input 


by 

Frank J. Romano 
President 

Graphic Arts Marketing Associates 
Salem, New Hampshire 



Produced in conjunction with 


National Composition Association 


Section of 



Printing Industries of America, Inc. 


GRAPHIC COMMUNICATIONS CENTER, 1730 NORTH LYNN STREET, ARLINGTON, VIRGINIA 22209 TELEPHONE: 703/527-6000 TELEX: 89-2738 


CABLE ADDRESS: PRINTINAM, ARLINGTON, VA. 




This book is dedicated to the memory of Arthur L. 
Koop, Jr., Manager of Advertising and Sales 
Promotion for Mergenthaler Linotype Company. 
Thanks Art. 


The bulk of typesetting for this book was done on Photon 713-5 
photo typesetters using the Warlock Random Access Composition Entry 
system for input. The card medium (covered in Chapter 5) permitted the same 
kind of correction flexibility that would have been available in hot metal slugs. 
The author wishes to thank Warlock, its president, Ernest Griesbach, and 
especially its pretty and efficient keyboard operators. It is also with great 
appreciation that thanks are extended to the many manufacturers that 
provided information and assistance; and also to Roy Mochi, and to Don 
Goldman. 



Table of contents 

1. In the beginning 

2. Telling a typesetter what to do 

3. Keyboard to tape systems 

4. Input media and coding 

5. Punched card input systems 

6. Video display terminal (VDT) systems 
. Optical character recognition (OCR) systems 

8. What a computer does 

9. Word processing 

10. Keyboard arrangements 


Page 1 
Page 9 
Page 35 
Page 55 
Page 63 
Page 71 
Page 85 
Page 107 
Page 117 
Page 139 



Early in 1926, Walter W. Morey suggested that equip¬ 
ment could be developed for automatic operation of 
linecasting machines from one or more remote points. 
This suggestion was discussed with Frank E. Gannett, 
owner of a chain of newspapers in the eastern U.S. 

Later that year, Gannett and Morey discussed this 
matter with Sterling Morton, Howard Krum, and Ed¬ 
ward Kleinschmidt, owners of what was then the 
Morkrum-Kleinschmidt Corporation, which manufac¬ 
tured most of the wire communication equipment in 
this country. The name of this corporation was later 
changed to the Teletype Corporation. After several 
conferences, it was agreed that Teletype would 
manufacture the equipment developed by Morey. 

In 1927, a Teletypesetter Perforator was ready - based 
on the principles used in the six-unit perforator then 
manufactured for communication purposes. The 
equipment was tried at the McCarty Typesetting 
Company in Chicago, and during the first day fifty 
five lines were set covering the Lindbergh flight to 
Paris. 

By 1928, the equipment was considered sufficiently 
well developed for prelininary unveiling to the field. 
Accordingly, a demonstration was arranged on the 
top floor of the Rochester Times Union Building (see 
photo to right). 

During 1929, a separate corporation was organized 
under the name of the Teletypesetter Corporation to 
handle this type of business, and from then to 1930 an 
experimental installation was made at the Evanston 
(Illinois) News Index. 

Manufacture of equipment was curtailed during the 
war. By 1951 the press associations began tran¬ 
smitting justified tape to newspapers and after one 
year about 400 daily newspapers were using this ser¬ 
vice. 

The use of Teletypesetter grew at a rapid pace. In 
1958 Fairchild Graphic Equipment purchased the 
business of Teletypesetter and in 1964 reported that 
over 1200 daily newspapers, 400 weekly newspapers 
and 350 commercial shops were using its tape equip¬ 
ment. 

In use today, there are about 15,000 Perforators and 
12,000 Operating Units. In 1972 Fairchild was sold 
to the Varityper division of Addressograph 
Multigraph. Below are two new versions of what 
began way back in 1926 when one man had an idea - 
an idea and an accomplishment that has to be con¬ 
sidered one of the milestones in the progress of the 
graphic arts industry. 















1. In the beginning 


Over four hundred years passed before Gutenberg’s movable type moved by 
machine rather than by hand. Even as late as the last decade of the 
Nineteenth Century and the tail-end of the Industrial Revolution a 
newspaper compositor set type almost the same way a Fifteenth Century 
printer did. From a large case in front of him he selected the necessary 
letters, numbers and other characters to form the line of type and then filled 
out the line with spaces to the required width. He did this over and over 
again for every line. A good worker could zip along at about 40 of today’s 
newspaper lines an hour. 

The development of the Linotype is a unique story all its own and this is not 
the place to retell it. Suffice it to say that Ottmar Mergenthaler, after several 
false starts, decided not to use foundry (or handset) type but rather to cast an 
entire line from matrices. One of his major problems was the automatic 
justification of the line of type which was solved by the sliding wedge or 
spaceband patent he obtained from J. W. Schuckers. Typesetters, on the 
average, set type by machine six times faster than by hand. Employers, hard- 
nosed businessmen that they were even then, sought to increase this 
profitability by utilizing unskilled operators, preferably women. However, 
the unions and experience showed that re-trained hand compositors were 
more appropriate. An active Fern-Lib movement in 1900 could have caused 

i | 

a different kind of revolution in the composing room. 

In 1887 Rolbert E. Lanston unveiled the Monotype. Although constructed 
differently in many ways from the Linotype, its major difference involved 
the non-direct operation of the machine. A perforated tape was prepared on 
one unit, and then run on another to cast the characters. Corrections were 
made without re-setting an entire line as with the Linotype. It is interesting 
to note, with 20-20 hindsight, that tape typesetting was born at the same time 
the Linotype was. 

In 1928 Walter Morey, a Monotype operator, invented the Teletypesetter. 
He experimented with tape as the typesetting medium to run a specially 
adapted Linotype, applying the principles of the Monotype to the dominant 
typesetting device of the day. 


Composition Input 1 



Typists, with minimal training, could produce tapes, which when fed into the 
Linotypes, produced metal slugs as though an operator was at the keyboard. 
In 1951 the Associated Press announced its TTS service and a year later 
United Press did the same. News now came into a newspaper as a tape which 
could activate a Linotype to produce type automatically. Billboard, a weekly 
newspaper, that year signed the first TTS union contract, outlining plans for 
perforation of tapes in New York for transmission by wire to its midwestern 
composing room. By 1953 almost 700 daily newspapers utilized TTS 
equipment, totalling about 1,400 keyboards for “punching” tape and about 
1,500 units for feeding tape to the Linotypes. 


Teletypesetter Corporation promoted its system by saying that a beginning 
typist could produce 400 or more lines of type an hour. In 1958 Fairchild 
Graphic Equipment, Inc. purchased the assets of Teletypesetter and 
introduced new models which offered greater type mixing capability and 
other productivity features. In 1971 it was reported that over 15,000 TTS 
keyboards were built, with many still in use. 

Since Linotype operators required retraining to use a TTS keyboard, union 
schools were established. The ITU even developed the Brewer Keyboard, 
invented by Claire N. Brewer, which was essentially a Linotype Keyboard 
designed to fit over a TTS keyboard. The operator now punched tape on a 
familiar keyboard arrangement. 


Operating a TTS keyboard was more fatiguing than typing. There were 
nineteen additional keys over the normal typewriter set; the operator had to 
move his eyes from copy to pointer regularly to monitor justification, plus 
scan the tape for possible errors. In spite of this, the major advantage of tape 
operation was not at the keyboard (at this point in time) but in the efficiency 
of Linotype operation. Three Linotypes running from tape could produce as 
much type as seven Linotypists. By 1959, almost 1,200 newspapers had 
installed TTS equipment. 



The greater portion of this editorial page was composed on the first 
commercial Linotype placed in a Newspaper Plant. It revolutionized com¬ 
posing room methods and made possible the great Newspapers of today. 



N. Y. Tribune Editorial Page, July 3, 1886 


A woodcut of Mergenthaler demonstrating his new 
machine. Whitelaw Reid (right) of the New York 

Tribune named the Linotype. The first newspaper page set on the Linotype. 


2 Composition Input 












































































































































































To convert the typesetting process from hand to machine a number of 
developments were necessary. The most important, and basic principle on 
which the linecaster rests, was that of producing lines of type to the same 
width. This was solved by the spaceband, a wedge-shaped device that 
expanded the space between words and forced each line to its maximum 
width. The spaceband was (and still is) a mechanical device with minimum 
and maximum expansion values; whereas in future systems the function of 
the spaceband would be (and was) imitated arithmetically. 

The Monotype System introduced tape operation long before the advent of 
TTS. Consisting of two units, the keyboard and the caster, Monotype 
computes what the linotype expanded. Like any successful system, 
Monotype rests on a basic principle. In this case it is the Monotype unit 
system. First of all, it is necessary to understand the steps that lead to 
justification of a line of type: 

1. The width value of every character is added, plus a minimum value for 
each word space. 

2. The resulting total from (1) is subtracted from the pre-determined 
line length. 

3. The number of word spaces is counted and this total is divided into the 
remainder left from (2). 

4. The amount of space determined by (3) is inserted at each word space. 

5. The line now equals the required measure. 


In formula form: 

Total line length less 
Total of characters and spaces 

Individual word space 

. .. .— = amount needed to 

XT , r , space line out 

Number ol word spaces 

to measure 


Thus it is extremely important that the width value of every character is 
known. This is the purpose of the unit system, a method that will be re¬ 
incarnated in future tape operation. 

Monotype matrices were designed with a maximum value of 18 units. This is 
a relative measurement, not an actual width value. The “W”, “M” and other 
characters are allotted the maximum of 18 units. Here is a chart indicating 
characters segregated according to a unit system. Designers may vary 
characters to fit a different unit value; thus, not all letters always occupy the 
same unit value, hut all must fit into one of the established unit categories. 


Character allotment in an eighteen unit proportional 
spacing system. 

Unit 

width Characters 

6 il.,'- 

7 jft 

8 I 

9 rsz01 234567890$; : ! 

10 c e o 

abdghnpquvxykJS 

12 Z 

13 CTL 

14 ABFOP&QV 

15 wDEGRUXYHKN 

M m M W % 

: i»cr j nit is exactly 1/18th of the width of the capital letter 


Composition Input 3 



As the Monotype Keyboard operator punched the special tape for activation 
of the caster, a mechanical wheel “added” the unit values of each character 
and space in a line. Word spaces per line were also counted and indication 
made of the need for an end-of-line decision. A scale informed the operator 
which keys were to be struck to obtain the correct word space width. The 
resultant tape was then fed into the caster tail-end first so that the spaces 
would be cast first. 

Impetus for the development of tape-run linecasting came in the 1930’s with 
the growth of newspaper groups and chains. The ability to share tapes 
among several papers, since certain types of articles of news were produced 
word for word in many different locations, seemed to be an answer to the 
problems of composition productivity. The existence of the nationwide 
Teletype communications system added fuel to the tape-selling fire. In 1933 
Walter Morey introduced Teletypesetter and the ability to input copy a 
single time on tape and transmit it to various locations by telegraphic 
methods. 

Tape perforation was a whole new ballgame. It was unlike linecasting; its 
keyboard was a standard “QWERTY” one type rather than the Linotype’s 
“ETAOIN,” and instead of monitoring the assembly of matrices, the 
operator monitored a width accumulation pointer to make the necessary 
end-of-line decisions. Since computational functions as to the number of 
characters, number of spacebands and their minimum width values were 
performed at the keyboard, the tape replaced the linecasting operator but 
continued to perform all of the linecaster’s functions. 

TTS was a transmission system in its original form. Tape generated at a 
central location was transmitted by wire to another location. There a 
duplicate tape was produced on a Reperforator. Linecasters were equipped 
with the Operating Units to read the tape and Adaptors to replace the 
keyboard. 

The TTS Perforator had a standard typewriter layout with the exception of 
the SHIFT key which did not work like a standard typewriter SHIFT key. 
To access a capital letter the operator hit SHIFT, then the character, and 
then UNSHIFT in order to return to the lower case. An important part of 
the Perforator was the counting mechanism, which accumulated (added) 
character and space values. TTS required a unit system somewhat similar to 
the Monotype Linecasting System. Matrices were made according to values 
established by dividing the EM space into eighteen parts. Eleven groups of 
width values, starting at 6 units (the narrowest) and going to 18 units (the 
widest), were used to divide up a font. 

The TTS Perforator produced no hard copy (that is, no typewritten end 
product). It was among the first “blind” keyboards and its only function was 
to create proper input tapes for a linecaster. The operator observed the scale 
in front of him and made end-of-line decisions (hyphenation, etc.) when the 
pointer entered the justification zone. Other keys on the perforator 
controlled linecaster machine functions. 

Since the unit system restricted TTS to newspapers, another version of the 
4 Composition Input 



perforator was developed to count non-unit matrices. The Multiface 
Perforator, as it was called, could handle those typefaces designed on a 32 
unit system. The EM was divided into 32 units with 5/32 of the EM as the 
narrowest character. There were 28 groups of width values. 

Because all typefaces vary in the widths of their characters, especially in 
fonts of different point sizes, TTS had to solve the problem of mixing. The 
answer came in the form of a width encoder, called a Counting Magazine by 
TTS, which recorded the width values for all characters in a particular 
typeface. From unit cut, to non-unit cut, to mixing, Teletypesetter evolved 
into a broadbased tape system for automatic linecaster operation. 


Photographic typesetting enters the picture 


In 1968 the direct input concept rose from the ashes 
in the form of Compugraphic’s 7200 display 
phototypesetter. It produced type from 14 to 72 point 
in any of four typefaces all under direct operator con¬ 
trol. 


There are differing opinions as to the birthday, and the father, of 
photographic typesetting. Practically, as well as diplomatically, there were 
several of both. In 1944 an ITT engineer in France named Rene A. Higonnet 
visited an offset printing plant in Lyon and observed the difficulty of 
adapting hot metal composition to the photo-lithographic process. Recalling 
a technical paper on the use of flash tubes to study high speed mechanisms 
such as propellers, he discussed an application of this principle with a friend, 
Louis Moyroud. By 1945 they had put together a simple machine to 
demonstrate that a flash was fast enough to project on a plate a clear image 
of a character from a revolving disc. The second problem, line justification, 
was solved by April 1946 in a unit that composed lines on film from a 
selection of 82 characters. The main components of the machine were: 



In May, 1946 Higonnet visited the U.S. and, attempting to interest 
investors, found his way to William W. Garth, Jr., President of Lithomat 
Corporation developers of lithographic plate materials. An agreement was 
reached and a second version was demonstrated in July 1948. This unit 
offered 88 characters, improved speed, and one of the first electric 
typewriters, an Electromatic. A variable escapement mechanism permitted a 
selection of 18 character widths. In 1949 the machine was shown to the 
industry. 


a. A manual typewriter with permutation bars and 
contacts to give each character a unique binary code. 

b. A memory register to record the codes on a 
revolving drum. Solenoid-controlled pins could be 
set for either of two positions. 

c. A counter-justifier to accumulate character widths 
and determine word space increments after each 
line had been typed. 

d. Control circuitry composed of telephone relays 
(ITT, remember?) 

e. A photographic unit made up of a film carriage 

and a continuously rotating disc with 82 clear characters 
on a black background (a photo positive). 


Composition Input 5 











This is the Photon 200B. It was the first true 
photographic typesetter to be introduced com¬ 
mercially. Like the Linotype it was controlled direc¬ 
tly by the operator, who, sitting at the keyboard, 
could set sixteen typefaces in twelve type sizes in a 
variety of typographic formats up to 72 points. 
Whether composition should be produced directly or 
indirectly (off-line tape) is still an area of concern to 
those who set type. 


In September, 1950, at the Chicago-held Graphic Arts Exposition, Intertype 
unveiled the Fotosetter, a machine that composed type on film or 
photosensitive paper, instead of hot metal slugs. To the naked eye the 
Fotosetter was a linecaster, except that the pot of molten lead was replaced 
by camera equipment. It contained 114 instead of 90 keys and used a matrix 
very similar in outward appearance to the casting matrix, again with the 
important exception of a photo-negative character embedded in its side. 
Characters could be enlarged or reduced from 6 to 36 point in a number of 
sizes in-between from the 12 point Fotomat at the turn of a dial. 
Complicated mixing of sizes and typefaces was greatly simplified over the 
hot process where a separate magazine was necessary for each size in each 
typeface. Mergenthaler, at this exhibition, showed an aborted version of the 
Linotype that used ebonite matrices (the line was set and the characters 
photographed). Linofilm, as it was called, went back to the drawing boards. 

The Fotosetter and the Higonnet-Moyroud machine continued the direct 
operation concept of the linecaster. 

Meanwhile back at the H-M unit, development continued. It was decided 
that the machine would operate from a standard typewriter keyboard, and 
that all controls should be at the keyboard position. Other decisions: 
justification should be from one typing, line lengths up to 42 picas 
(Linotypes used a standard 30 picas; 42 pica units were more money), line 
length should be under dial control; typefaces should be mixed in the same 
line, as well as type sizes, lines should be able to be “Killed”, and all 
quadding functions should be automatic. They were far from achieving these 
goals. So more money was raised and Lithomat changed its name to Photon. 

By 1955 ten machines had been installed. Time for development of the 
expanded unit allowed time for other competing units to reach fruition. The 
Linofilm from Mergenthaler (now a true photographic device using a grid 
for master images) and the Monophoto from Monotype were announced. 
The Photon unit was way ahead in terms of capability, offering sixteen fonts 
in twelve different sizes. In late 1955 further improvements were made and 
the 200 series machines were introduced. 


6 Composition Input 





The year 1956 saw the following inventory of industry phototypesetting 
devices: 



Fotosetter 

Photon 

Monophoto 

Linofilm 

ATF 


250 since 1950 
62 since 1950 
6 since 1955 
2 since 1955 

31 since 1954 


about $33,000 each 
about $55,000 each 
about $35,000 each 
about $57,000 for one 
keyboard , one photo unit 
about $14,000 each 


The American Type Founders unit incorporated the Flexowriter keyboard 
to prepare tape to drive a photographic output unit. In 1951 the Friden 
Company introduced the Justowriter, a two-unit direct impression 
typesetting system. An operator typed his copy on one electric typewriter 
(Recorder) perforating a 7-level paper tape, at the same time. Codes on the 
tape for characters and justification were needed for the output unit 
(Reproducer) which typed justified lines. ATF replaced the output unit with 
a true photographic, instead of strike-on, device. The Monophoto unit 
retained the tape principle used in its casting units. Linofilm also advocated 
the separation of keyboarding and photographing. 

Aware of the tremendous growth in the use of punched-tape from about 1950 
to 1960, Photon decided to develop a tape capability into the 200 series 
machine. Two models were planned; one to accept TTS tape (the 510 series) 
and one to accept Monotype tape (the 520 series). Of course, users were 
restricted in accessing all capabilities of the Photon photo unit because of the 
inherent limitations of the TTS and Monotype code structures. To solve this 
problem and better meet the competition of the multiple keyboards to one 
photo unit concept, a new series was begun. The 540 series used the photo 
unit and keyboard from the 200 series but separated them. In December 
1962, 12 photo units and 5 keyboards were shipped. The keyboards 
perforated 8-channel paper tape. 

By 1964 new photographic typesetters were being readied. Between then and 
1967, the Linofilm Quick, Fairchild Phototextsetters, Photon 713 series and 
Intertype Fototronic were introduced. Most accepted 6-level TTS paper 
tape, by now the dominant input medium. In 1968 Compugraphic 
Corporation introduced the forerunners of a phototypesetting line priced 
below all others in the industry. TTS tape was the input. Many (and that 
means many) other phototypesetters were introduced between 1968 and the 
present. A survey of them and their capabilities is included later in this book. 

As can be seen, typesetting and input evolved along parallel, but separate 
paths. 


Composition Input 7 




Mergenthaler’s Linofilm was introduced shortly after 
the Photon 200B. It separated the functions of 
keyboarding and output and was the first true 
phototypesetter to demonstrate the principle. Two to 
three keyboard operators could keep one photo unit 
busy. 


The Elektron was Mergenthaler’s new breed of 
Linotype in the early sixties. It was designed 
specifically for tape operation and could zip along at 
fifteen newspaper lines per minute. The Intertype 
Monarch was also designed specifically for tape, and, 
in fact, was introduced and sold without a keyboard. 




position Input 











2. Telling a typesetter what to do 


There was a time when typesetting meant only the process of assembling 
individual pieces of type by hand. Gutenberg printed his Bible from movable 
type in the middle of the Fifteenth Century in this way. The first major 
advance in typesetting came centuries later with the invention of the casting 
typesetters. Another major advance in typesetting occurred effectively 
within the last five years with the development of photographic typesetters. 

Typesetting cannot be easily discussed without reference to hand typesetting 
and to the operations of Linotype and Monotype machines. This is so 
because many terms used in typesetting have no literal significance except 
with respect to the earlier processes — since it was with these earlier 
processes that the terms originated. 

It is the purpose of this section to present some basic facts about type 
composition. All of the important terms will be introduced. 

An understanding of typesetting requires having at least an elementary 
knowledge of typography. The most basic element of this knowledge is in 
knowing what the various printing characters are called. 

Printing characters 

In the English alphabet there are twenty-six letters with which words are 
formed. Each letter appears in several kinds of printing characters: upper 
case letters, lower case letters, small capitals, initial letters, ligatures, and 
logotypes . 

Upper case letters are called capital letters since they are used for satisfying 
the rules of grammar which requires that the first letter in every sentence be 
a capital and that proper names be capitalized. 

The terms upper case and lower case are derived from the arrangement of 
cases used by printers for holding type. Capital letters are distributed in the 
upper case; hence, the term upper case letters. 


Composition Input 9 



Small capitals are similar to upper case letters but are designed somewhat 
differently in order to form entirely capitalized words having a pleasing 
appearance. The height of small capitals is always somewhat reduced from 
that of upper case letters. 

Initial letters are ornamental or large capital letters used to embellish 
composition. 

A ligature is formed by combining two or more letters into a unique design. 
Early printers, following the practices of the scribes, had many more 
ligatures than is now customary. In book composition and in many 
periodicals, but not newspapers, the f-ligatures are common. 

Ligatures for the dipthongs ae and pe are familiar printing characters for 
both upper and lower case letters and for small capitals. 

A logotype , defined with respect to handset type and Monotype, is a 
combination of printing characters cast from a common matrix. The 
simplest kind of logotype is a combination of letters spaced in a way that is 
not possible with ordinary type. For example, the italic letters Y and e 
spaced according to the physical requirements of ordinary type result in an 
unattractive gap; the same letters, with improved spacing, look much better 
as a logotype. Because there is no inherent limitation in photographic 
typesetting in achieving any desired spacing, the provision for generating 
logotypes is simplified. Modern phototypesetters utilize a “kern” command 
to “back up” and tuck characters together. 

Many logotypes are distinctive combinations of letters and other printing 
characters. Ligatures can be included within the definition of a logotype. 

Accented letters are part of many alphabets: French, German, Italian, 
Spanish, any foreign language. 

Figure is the term used by printers when referring to the numerals 0, 1,2, 3, 
4, 5, 6, 7, 8, and 9. There are two distinct classes of figures. Modern figures 
are designed to have a uniform height and sit on the base line of the alphabet 
to which it belongs. Old style figures are designed so that some figures 
descend below the base line and the figures do not all have the same height. 

Fractions are closely associated with figures. An extensive variety of 
fractions is available in type. 

Points, or the marks of punctuation include the following: 

, Comma 
; Semi-colon 
: Colon 
— Dash 
Hyphen 

4 Beginning quotation marks 
’ Ending quotation marks 


10 Composition Input 



[ Beginning brackets 
] Ending brackets 
(Beginning parenthesis 
. Period 

? Question Mark 
! Exclamation Point 
) Ending parenthesis 

The quotation marks are used simply or doubly. The apostrophe is of course 
identical as a printing character to a single ending quotation mark. 

The ampersand is a mark which represents the word and. There are common 
and italic forms of the ampersand. 

The marks of reference include the following: 

* Asterisk 
t Dagger 
$ Double Dagger 
§ Double-S 

The paragraph mark can take many forms. 

The following commercial and monetary signs are familiar printing 
characters: 


% 

Per Cent 

lb 

Pound 

# 

Number 

c/o 

Care Of 

$ 

Dollar 

£ 

Sterling 


Superior and inferior characters include both letters and figures and are used 
in mathematical work. Superior figures are used to denote references in 
scholarly texts. 


Special characters are used in printed matter for a wide variety of purposes: 
ornamental devices, arrows, astronomical marks, mathematical symbols, 
ecclesiastical signs, musical notations, bullets, ballot boxes, etc. 

Horizontal and vertical rules are used to form tables. A careful examination 
will show that each vertical line is composed from a series of short rules. 

Leaders are dots used in many types of composition, such as telephone 
directory listings. 

An EM leader has two dots; an EN leader has one dot. 

The term typeface originated with hand typesetting. A typeface is the flat 

Composition Input 11 



surface at the top of a piece of type; when subsequently inked and pressed 
against paper, an impression of the printing character is obtained. The face 
is thus a reverse image of the printing character. 

The term face has broad significance in that it is used in describing styles of 

type. Thus one can generalize about bold face types or become specific and 
refer to Caslon Old Face. 


A typeface may have a number of features which have an effect on spacing 
and on how well characters are reproduced. A stem is a thick vertical line 
used in forming a letter; a hair line is a thin one. Some letters consist only of 
straight lines and others, like the p consist also of loops or swells. A counter 
is a space surrounded by lines. A serif is a short cross-line at the extremity of 
a letter and is one of the most important characteristics of type style. Sans 
serif characters have no serifs. 

The middle parts of letters used in the same composition must all sit on a 
common base line. The part of a letter that drops below the base line is called 
a descender, the part that rises above the middle part is called an ascender. 
The entire vertical distance in which a type face is to be located is called the 
body height. The corresponding horizontal distance is called the width. 




The face of a piece of hand-set type is produced by casting. A matrix for the 
face is inserted into a mold. Molten metal is poured into the mold; the chief 
ingredient of the alloy is usually lead. Hand-set type is often referred to as 
foundry type because it is produced in foundries. 


The term matrix is directly applicable to both Linotype and Monotype 
casting operation. A Monotype matrix is used to cast at a given time a single 
piece of type quite similar to foundry type. A Linotype matrix is also called a 
mat. Linotype matrices are assembled as shown in the figure at right and a 
complete line of type is then cast in the Linotype mold. The resulting line of 
type is called a slug. 

Type is produced in photographic typesetting by projecting images onto 
photographic film or paper. Images are created by transmitting light 
through clear areas on an otherwise opaque film, drum, or disc. The 
characters formed by the clear areas may be called font characters. Font 
characters may also be generated using cathode ray tubes. 



Linotype lead slugs. The ability to access only a por¬ 
tion of typeset material for correction is one of the at¬ 
tributes of hot metal composition that photographic 
methods have been hard put to emulate. 


Font characters are analogous to the matrices of the type casting processes 
with an important exception. The correspondence between a matrix and the 
resulting typeface is one-to-one; however, the typeface produced from a font 
character need not be the same size as the font character. In the 
Compugraphic 7200 Phototypesetter, for example, a single font character 
can produce typefaces in eight different sizes by selecting from eight 
projection lenses. 


Type measurements 

The height, width, and spacing of type is measured in points and picas. There 
are three principal point systems in use today. The American-British System 


12 Composition Input 




is in use in English speaking countries. The point in this system is 0.01383 
inch; twelve points makes a pica. The Didot System is in use in France and in 
most countries of Continental Europe — but not in Belgium. The basic unit 
of the Didot System is called the corps or point ; it has a value of 0.01483 
inch. Twelve corps equal a Cicero. In the Mediaan System of Belgium, the 
corps or point has a value of 0.01324 inch; twelve of these equals a Mediaan 
and/or Cicero. 

Two measurements determine the size of an individual piece of type. 
Referring back, these are the body height and width. For foundry and 
Monotype, the measurements can be made on an actual piece of type. Type 
produced photographically, however, does not have any apparent boundries; 
but the boundaries, nonetheless, are implicit. The artwork from which 
photographic characters is made is drawn within an area having an outline 
corresponding to the borders of an equivalent piece of foundry type. 


When referring to a complete alphabet rather than to an individual piece of 
type, the alphabet size is defined by body height and set. Body height is 
always measured in points. 

Point size always refers to the body height. Thus, 10-point Scotch Roman 
specifies a Scotch Roman type having a 10-point body height. 

The width of an individual piece of foundry type is determined by the design 
of the printing character. For example, a 7-point lower case letter i would 
ordinarily be much narrower than a 7-point upper case B in the same 
alphabet. 

The width allotted a Linotype face is the same as the width of the Linotype 
matrix. Since Linotype matrices are made from brass, the term brass width 
is often used to denote the width of a Linotype character. 

The set of an alphabet is a relative measure of the width of the entire 
alphabet. The em space is divided into a number of relative units (RU). The 
width in points of the em is the set of the alphabet. Every character in the 
alphabet is designed so that its width is an integral number of relative units. 
The actual width of a character in points is then 



I-ARACTERS 

Matrices and spacebands assembled and ready for 

casting. 


RU in character 


x Set 


RU in em 

where set is in points. In most typesetters the alphabets are designed with an 
em having 18 relative units; Justowriter alphabets are designed with an em 
having 5 relative units. 


For foundry type, an em is a square space having a width equal to the body 
height. 

The body height of type produced on a phototypesetter is determined by the 
size of the font character and by the magnification of the lens used to project 


Composition Input 13 




PROJECTION LENS 


FONT CHARACTER 


an image of the font character. The Compugraphic 2900 and 4900 Text 
Phototypesetters have two lenses so that the type size is controlled by the size 
of the font character (same size lens) or twice size (2X lens). In the CG 7200 
Headline Phototypesetter the font character is always 10 point. By means of 
eight projection lenses, type from 14 points to 72 points can be produced 
from the same 10 point font characters. 

When a character has been photographed, the width stepping motor will 
take as many steps as there have been relative units assigned to the 
character. The motion of the width stepping motor is modified by the set 
gears in such a way that the lens carriage will move so that the optical axis of 
the system will shift an amount equal to the width of the type. Different sets 
are provided for by changing set gears. 



LENS 

CARRI 


WIDTH STEPPING MOTOR 


Font & font strips 

A font of type consists of an assortment of printing characters in a size and 
style. Thus, ten point Electra is the description of a particular font. The 
exact assortment of type in a font will vary somewhat depending on the 
requirements of composition. 


A Standard Roman font includes upper and lower case letters, small 
capitals, figures, marks of punctuation and reference. Some fonts may 
contain ligatures or fractions, and others may not. 

A font of foundry type is stored in the printers’ cases. There is a separate 
partition for each character. A character for which there is no partition is 
called a pi character. When two different fonts of foundry type are used in 
the same composition, the compositor simply goes from one pair of printers’ 
cases to another to collect the pieces of type required. 

In Linotype operation, two or more fonts are available, depending on the 
capabilities of the particular machine in consideration. This is accomplished 
by two operations: the rail shift and the magazine shift. 

Some Linotype machines are equipped with two magazines — an upper 
magazine and a lower magazine. Some Linotypes have as many as four 
magazines; these are usually designated as MAG-1, MAG-2, etc. Each 
magazine stores an assortment of ninety different matrices. A matrix that is 
not stored in a magazine is called a pi mat or pi matrix and is said to run pi. 
Referring back to the figure of the Linotype matrix, it is seen that a Linotype 
matrix contains two different faces. When a matrix is assembled, either one 
or the other of the faces is aligned in the casting position. This is 
accomplished by the rail shift. A character is thus assembled in the lower rail 
position, or upper rail position. By means of the rail shift, companion faces 
can be composed in the same line. 

The concept of upper and lower rails, and upper and lower magazines is 
applied to the identification of the style choices available. A font strip for a 
CG 2900 Series Phototypesetter has a single track of characters. The first 
half of the font strip contains the lower rail style, and the second half the 
upper rail. The font drum of a CG 7200 Phototypesetter can mount two 


A two-font film strip. Upper half for lower rail; lower 
half for upper rail. 



LOWER 

RAIL 

FONT 


UPPER 

RAIL 

FONT 


LOWER 



UPPER 

RAIL 


LOWER 

MAGAZINE 


UPPER 

MAGAZINE 


A four font film strip. Now the upper half is the lower 
magazine and the outside row the upper rail. 


14 Composition Input 





UPPER MAGAZINE 
FONT STRIP 


Film strips are mounted on drums with position 
nomenclature the same as that used on the linecaster. 


The Photon glass matrix disc. 


LOWER MAGAZINE 
FONT STRIP 


separate font strips. One font strip is mounted in the lower magazine (LM) 
position, and the other in the upper magazine (UM) position. In addition, 
each font strip has two tracks, one of which is designated as being in the 
lower rail, and the other in the upper. 


- .recasting magazine. It has 90 channels, each for a 
raaque character - in one point size. 



The term composition when applied to printed matter, means the selection 
of type by size and style, and its arrangement. Even a cursory survey of 
newspapers, books, circulars, calling cards, posters, and pamphlets will give 
some indication of the great variety possible with printed matter. 

The size and style of type and its spacing are selected to satisfy readability 
and appearance. The general appearance of a column or page of type will 
create a distinctive light or dark appearance. This appearance is referred to 
as the color of type. In photographic type compositions color is determined 
not only by the selection of the face, but by the photographic process. The 
control of photographic density is thus very important in achieving good 
typography. 

Straight matter is simple justified text. 

Except for the paragraph indention of the first line, every line is flush with 
the left margin ; except for the last line, every line is flush with the right 
margin. The length of a line is measured in picas and points. Type not flush 
with the left or right margin is ragged. 


i'E'R RAIL CHARACTER 


ER RAIL CHARACTER 



A linecasting matrix. 


A line of foundry type is justified by filling the line with words; or, following 
the rules of hyphenation, by ending the line with part of a word and a 
hyphen. 

To fill out a line so that it is flush with the right margin, spaces are inserted 
between words called interword spaces. In Linotype composition, the 
interword spaces are filled out by inserting space bands between words. 
Space bands are double wedge devices that expand in thickness when 
mechanically acted upon. 


Composition Input 15 




In some cases it is necessary to add spaces between the letters of a word - a 
process known as letterspacing. 

Justification in computer-controlled photographic typesetting is effected by 
calculation. The space required to fill a line is divided into the number of 
interword spaces and, if necessary, one or more words are letterspaced. 

Sometimes the spaces in a number of lines of type cause noticeable white- 
connected areas called rivers. 

The various spaces used in composing a line have names related to their 
relative thickness. An em space in photocomposition is a space having a 
width based on the full number of units used in designing the type. Thus for 
an 18 unit alphabet, the em has eighteen relative units. The actual width, of 
course, depends on the set of the type as previously explained. In foundry 
type, the em is a square of the type body. A square space is also called a 
quad. The term quadding applies to filling out a line with space. An en space 
has half the thickness of an em space. The space used in letterspacing is often 
called a thin space or a hair space. 

. The spacebands are pushed up, or “expand” to fill the 

When composing foundry type, the spacing of characters is determined by word s P aces - 

the width of type and the use of additional spaces. In photocomposition, 

spacing is determined by the distance one photographic image is offset from 

another. This offset is measured from the vertical reference line of one 

character to the vertical reference of the following one. The term escapement 

is often applied to this width because some early phototypesetters used 

escapement mechanisms for obtaining character displacement. 



In some composition it is desirable to place type closer together than is 
possible in the case of foundry type if kerning were not used. Kerning allows 
part of a character to extend into the body of an adjacent character. Kerning 
is, of course, easily accomplished in photographic typesetting. 

The terms flush left, flush right, and center are used to describe unjustified 
lines flush with the left margin, the right margin and centered. Display 
composition, the type of composition used in advertisements, utilizes flush 
left, flush right, and centered lines. The term Quad is synonymous with 
Flush). 



The Linotron 505 was introduced by Mergenthaler in 
late 1967. It utilizes a cathode ray tube (CRT) for 
composition and originally required input via a com¬ 
puter. The present version, 505C, has its own com¬ 
puter and can thus accept unjustified tape. Mergen¬ 
thaler has a special keyboard designed specifically for 
the 505. 


16 Composition Input 








The space between lines of type is controlled in foundry type by the body 
height of the type and by the addition of leading between lines. Leading is 
strips of lead, hence the term. In photographic type composition, the spacing 
is determined by the amount by which the base line of a line of type is 
displaced from the baseline of the preceding line. The term leading when 
applied to photocomposition means the displacement of lines of type. 
Leading is usually measured in points. 


Justification 

Here’s how a line of type is justified in a “computer” system or stand-alone 
typesetter with “computer” capability: 


All typesetting starts with initialization. One must define the parameters and 
format of the material to be set in type: point size, type style, leading, and 
line measure. This is also called “dressing” the typesetter. 


S*t Solid 



As characters are keyed (input directly or input via tape) they enter the 
computer’s or typesetter’s logic under the same kind of scrutiny that dance 
hall girls come on stage. The width value of every character is deducted from 
a counter on which the total line measure has been recorded. 


Characters are, of course, read at speeds of a few to a few hundred a second. 
The typesetter continues to deduct width values and also counts the 





Certain letter combinations may be examined in the more sophisticated 
computer system, such as an “ff ’ combination. Here the computer may be 
programmed to replace this letter set with a ligature for more aesthetic 
typography. Also, this combination is noted in case a line ending must be 
made. Characters continue to be deducted. 



By now the end of the line is approaching. Width deduction is approaching 
the “justification” zone. This zone is determined by a calculation that counts 
the number of word spaces and establishes minimum and maximum 
expansion values. If a word space cannot expand, the line is too tight; if it 
can expand too far, the line is too loose. 




18 Composition Input 


All characters that may possibly fit on a line are “in”. However, if the line is 
too tight, a letter or several letters must be removed. There are three 
alternatives: drop the word to the next line, or hyphenate it and drop certain 
characters to the next line, or force justify the line. 

First, hyphenation is tried. There are two basic ways: by rules of logic, 
grammatical constants that will work in a majority of cases, or dictionary, a 
look-up system where words are compared to a separate word list with 
acceptable break points indicated. 



One character at a time is dropped until a good hyphenation point is 
reached. 



If it isn’t, the machine will attempt to keep all characters on the line with 
word spaces shrunk to their absolute minimum width. 






We have now input all our characters and determined our line ending point. 
If our measure was set at 30 picas and all characters total 22 picas, the eight 
pica difference is divided into the number of word spaces. In this example 
there are eight word spaces, so each space is one pica. If the word spaces 
were too wide, thin or hair spaces would be inserted equally between all 
characters. This is called letterspacing. 


Measure 
Total character width 
Total interword space 
Number of spaces 


divide 


30 picas 
22 picas 


8 

8 


picas 



1 pica per interword space 


The final line is now photographed, the same total width (measure) as every 
other line. All left and right margins line up. This is justification, one of the 
most important functions performed in typesetting. 



And by the way, the entire sequence described, although it may vary among 
typesetters or computers, takes something less than one quarter second to 
occur. 



second 


20 Composition Input 



How to drive a typesetter 


The preceeding information has laid the groundwork for an understanding of 
what a typesetter does. To review 
A typesetter: 

a. Produces various typographic characters 

b. Produces various sizes of characters 

c. Produces various styles of characters 

d. Produces various line lengths of characters 

e. Produces various placements of characters 
(e.g.: quadding, leading) 


Thus an input device must provide the methodology to access the specific 
characters and capabilities of which the typesetter is capable. The 
illustration on this page is a list of all the flashable characters in one font of a 
102-character phototypesetting font. Thus the input device must have 
sufficient keytop designations to allow the operator to locate and set these 
characters. The input device must also have the ability to access all or some 
of the following typesetter commands: 


ili» nnfsi£Bnation. . . 
"• fl ; mw - nrfiguration 


CASE CHAR 

(Shift) ACTER 


5 4 3 2 1 0 

0 1 2 3 4 5 6 7 




J 


III 


n 


*11 

■11 

* 

mu 

■ ill.; 

Ilf' 

!i" 

H 


0 0 0 11 0 0 

0 0 0 11 1 0 

110 0 1 0 0 

110 0 1 1 0 

0 1110 0 0 

0 1110 1 0 

0 10 0 1 0 0 

0 10 0 1 1 0 

0 0 0 0 1 0 0 

0 0 0 0 1 1 0 

0 110 1 0 0 

0 1 I 0 I 1 0 

110 10 0 0 

110 10 1 0 

10 10 0 0 0 

10 10 0 1 0 

0 0 110 0 0 

0 0 110 1 0 

0 10 11 0 0 

0 10 11 1 0 

0 I 1 I 1 0 0 

0 1111 1 0 

10 0 10 0 0 

I 0 0 I 0 1 0 

1110 0 0 0 

1110 0 I 0 

0 110 0 0 0 

0 110 0 1 0 

110 0 0 0 0 

110 0 0 1 0 

10 110 0 0 

10 110 1 0 

10 111 0 0 

10 111 1 0 


1 


1. 

2 . 

3. 

4. 

5. 

6. 

7. 

8 . 

9. 

10 . 
11 . 
12 . 
13. 


Quad Right 
Quad Left 
Quad Center 
Insert Space 
Insert Leader 
Insert Rule 
Tab Set 
Tab 

Back or Foward One Unit 
Back or Foward One Point 
One-half Point Line Space 
One Point Line Space 
Line Measure Set 

CASE CHAR- 

(Shift) ACTER 


14. Line Space Set 

15. Point Size Set 

16. Type Style Set 

17. Space Only 

18. Flash Only 

19. Carriage Reset 

20. End Line 

21. Super Shift 

22. Discretionary Flyphen 

23. Cancel Word 

24. Cancel Line 

25. Cancel Character 

CASE CHAR- 
(Shift) ACTER 


TTS Bit Configuration. 5 4 3 2 1 0 TTS Bit Configuration. .. 

ACM Code Configuration. 0 1 2 3 4 5 6 7 ACM Code Configuration 

CHARACTER CHARACTER 


5 4 3 2 1 0 

0 1 2 3 4 5 6 7 


r 

R 

s 

S 

t 

T 

u 

U 


V 

w 

W 

x 

X 


Y . 

z . 

Z . 

Base Line Rule (BLR) 

Star. 

Asterisk. 

Check. 

FRACTIONS 

Three-Quarters. 

One-Quarter. 

One-Half. 

One-third. 

NUMERICS 

1 . 

Superior 1. 

2 . 

Superior 2. 

3 . 

Superior 3. 

4 . 

Superior 4. 


0 

1 

0 

1 

0 

0 

0 1 

5 . 

. 1 

0 

0 

0 

0 

0 

1 


0 

I 

0 

1 

0 

1 

0 1 

Superior 5. 

. 1 

0 

0 

0 

0 

1 

1 


0 

0 

1 

0 

1 

0 

0 1 

6 . 

. 1 

0 

1 

0 

1 

0 

1 


0 

0 

1 

0 

1 

1 

0 1 

Superior 6. 

. 1 

0 

1 

0 

1 

1 

1 


1 

0 

0 

0 

0 

0 

0 1 

7 . 

. 0 

0 

1 

1 

1 

0 

1 


1 

0 

0 

0 

0 

1 

0 1 

Superior 7. 

. 0 

0 

1 

1 

1 

1 

1 


0 

0 

1 

1 

1 

0 

0 1 

8 . 

. 0 

0 

1 

1 

0 

0 

1 


0 

0 

1 

1 

1 

1 

0 1 

Superior 8. 

. 0 

0 

1 

1 

0 

1 

1 


1 

1 

1 

1 

0 

0 

0 1 

9 . 

. 1 

1 

0 

0 

0 

0 

1 


1 

1 

1 

1 

0 

1 

0 1 

Superior 9. 

. 1 

1 

0 

0 

0 

1 

I 


1 

0 

0 

1 

1 

0 

0 I 

0 . 

. 1 

0 

1 

1 

0 

0 

1 


1 

0 

0 

1 

1 

1 

0 1 

Superior 0. 

. 1 

0 

1 

1 

0 

1 

1 


1 

1 

1 

0 

1 

0 

0 1 










1 

1 

1 

0 

1 

1 

0 1 










1 

0 

1 

0 

1 

0 

0 1 

PUNCTUATION 









1 

0 

1 

0 

1 

1 

0 1 

Semi Colon. 

. 1 

1 

0 

1 

0 

0 

1 

1 

1 

0 

0 

0 

1 

0 

0 1 

Colon. 

. 1 

1 

0 

1 

0 

1 

1 

1 

1 

0 

0 

0 

1 

1 

0 1 

Close Paren. 

. 1 

0 

0 

0 

1 

0 

1 

1 

0 

0 

0 

1 

0 

0 

0 1 

Open Paren. 

. 1 

0 

0 

0 

1 

1 

1 

1 

0 

0 

0 

1 

0 

1 

0 1 

En Bullet. 

. 1 

1 

1 

I 

1 

0 

1 

1 

0 

1 

0 

0 

0 

0 

0 1 

Register Mark. 

. 1 

1 

1 

1 

1 

1 

1 

1 

0 

1 

0 

0 

0 

1 

0 1 

Hyphen. 

. 0 

1 

0 

0 

1 

0 

1 

1 








Dash. 

. 0 

1 

0 

0 

1 

1 

1 

1 

0 

1 

0 

1 

1 

0 

1 1 

Percent. 

. 1 

0 

1 

1 

1 

0 

1 

1 

0 

1 

0 

1 

1 

1 

1 1 

Superior $. 

. 1 

0 

1 

1 

1 

1 

1 

1 

0 

0 

1 

0 

0 

0 

1 1 

Slash. 

. I 

0 

0 

1 

0 

0 

1 

1 

0 

0 

1 

0 

0 

1 

1 1 

Superior Cent. 

. 1 

0 

0 

1 

0 

1 

I 

1 








Query. 

. 1 

0 

I 

0 

0 

0 

1 

1 








Ampersand. 

. 1 

0 

1 

0 

0 

1 

1 

1 

1 

1 

1 

0 

1 

0 

1 1 

Comma. 

. 0 

1 

1 

0 

0 

0 

1 

1 

1 

1 

1 

0 

1 

1 

1 1 

Comma. 

. 0 

1 

1 

0 

0 

1 

1 

1 

1 

0 

0 

1 

1 

0 

1 1 

Period. 

. 1 

1 

1 

0 

0 

0 

1 

1 

1 

0 

0 

1 

1 

1 

1 1 

Superior Period. 

. 1 

1 

1 

0 

0 

1 

1 

1 

0 

0 

0 

0 

1 

0 

1 1 

Close Quote. 

. 0 

1 

0 

0 

0 

0 

I 

1 

0 

0 

0 

0 

1 

1 

1 1 

Open Quote. 

. 0 

1 

0 

0 

0 

1 

1 

1 

0 

1 

0 

1 

0 

0 

1 1 

Dollar ($). 

. 0 

0 

0 

1 

1 

0 

1 

1 

0 

1 

0 

1 

0 

1 

1 1 

Exclamation. 

. 0 

0 

0 

1 

1 

1 

1 

1 


Characters that can actually be photographed include 
these above (ACM 9000). 


Composition Input 21 























































































































These tell the typesetter where and how to put the characters on the output 
medium. It is probably evident by now that an input keyboard, attempting to 
access all characters and functions via single or double keystrokes, would 
require a bewildering array of keys. Here is a photo and layout of 
Compugraphic’s ACM 9000 Keyboard. Note how it attempts to deal with 
single keystroke command of type style and size, two ever-changing areas of 
display composition. Single strokes here mean faster keyboarding. However, 
to access type style and size changes for the ACM 9000 via a standard 6-level 
keyboard requires a minimum of three keystrokes. To call in type style on 
Row 1, the operator keys: SUPERSHIFT, “F” (for face), “1” (for Row 1). 
Let us review why multiple keystrokes are needed. 




The ACM 9000 direct entry keyboard. 


The standard typewriter keyboard layout has 44 character keys (including 
SPACE); each key usually controlling two characters, which are selected via 
the SHIFT key. There are 26 alphabetic keys (for capitals and lower case 
letters), 10 numerical keys (for figures and special characters), 5 punctuation 
keys (but only 8 unique characters out of 10 possible positions, since the 
period and comma are repeated in the SHIFT position), and 2 extra keys for 
fractions and rule line. 84 unique characters from 88 character positions. 
The SHIFT and RETURN keys are function keys since they control the 
way the machine operates. 


22 Composition Input 




















The teletype has only 32 keys to access 64 character or function positions. 
There are only capital characters. 

The Linotype Keyboard layout has a total of 90 keys (plus the ELEVATE) 
to access the 90 channels of the magazine. Character matrices in each 
channel contain a duplexed pair of characters; that is, there are two separate 
positions, called UPPER RAIL and LOWER RAIL, both of which are the 
same width value. Thus the linotype has 90 keys to access 180 characters. 
This represents two complete alphabets in one type size and style, duplexing 
a typeface with its companion italic or boldface version. Note that the layout 
is arranged with the more frequently used characters more prominently 
positioned. Shown is keyboard Diagram 12. 



Along came the TTS Standard Perforator. It combined the standard 
typewriter layout with the function controls necessary to run the linotype. 
Compare the character set from both layouts. Missing are the small caps, 
since they were not used for newspaper work. The Linotype keyboard 
Number 282 the newspaper version (282 f included fractions). Unlike the 
typewriter, the TTS keyboard clearly differentiated between SHIFT and 
UNSHIFT. These were unique codes. The RUB OUT was a correction key: 
the tape was backed up to the incorrect code and when struck, RUB OUT 
inserted holes in all channels. UR and LR accessed the Upper Rail and 
Lower Rail position of the matrix. The BELL key signaled the receiving end 
of a TTS transmission that a tape was coming in. Other versions of the TTS 
keyboard were available to set non-unit cut typefaces, and to mix other faces 
and sizes by commanding UM and LM (Upper Magazine and Lower 
Magazine) changes. Keytops were also changed to represent characters 
running in the Linotype channels that deviated from standard layouts. 

As phototypesetting came more into use the number of characters per font 
increased. Sizes were changed by positioning lenses instead of entirely new 
fonts. More typefaces were available at one time. And electronically- 
controlled photographic typesetting provided more capability and thus more 
commands were necessary to access them. Intent on utilizing existing 
keyboards to drive some of the new typesetters, users were forced to strike 
more than one key to change a size or face or even flash one of the extra 
characters over the normal keyboard set. 

Subsequent keyboards increased the character and command set on the 
keyboard and, in some cases, provided multi-code keys. These keys 
perforated two or three codes in sequence with one keystroke. Two trends 
have gone foward simultaneously since then: 1. to adapt phototypesetters to 
run from the existing reservoir of keyboards, and 2. to design keyboards 
specifically to run phototypesetters. The result has not been one of logical 
order and method. 


This is Compugraphic’s 2961. It was the first 
phototypesetter ever priced under $10,000. For the 
price a user received a two-font typesetter that could 
accept justified and unjustified six level tape. A hard 
wired computer, based somewhat on CG’s Justape, 
provided hyphenation and justification. 


Composition Input 23 











Phototypesetting is in its third generation. First generation devices utilized 
operating principles from hot metal linecasting machines but used 
photography in place of actual metal casting. 

Second generation phototypesetters use photomechanical methods to select 
and expose character images on photographic material. These machines 
were developed in the 1950’s and have undergone continual refinement ever 
since. Third generation devices are fully electronic machines based on 
cathode ray tube (CRT) technology. They were developed around 1965 and 
are still undergoing significant changes. 


Second generation photopypesetters produce typeset copy from master 
characters stored photographically within the machine. The major 
subsystems used to produce the copy include input, character selection, 
image output control, and interline separation. Although the basic principles 
used by all manufacturers are similar, numerous variations in design 
complexity and configuration present a vast array of alternative methods for 
setting copy. 

Phototypesetters use codes from an input source to activate machine 
responses that produce typeset copy. The codes identify parameters that 
control the operation including: (1) characters to be typeset, (2) desired fonts 
and point sizes, (3) interword spacing for line justification, and (4) desired 
interline spacing. 



Linofilm Quick was supposed to be a low cost 
phototypesetter when it made its debut in 1964. It was 
priced around twenty thousand dollars and competed 
successfully with the sixty grand units then 
dominating the marketplace. The Quick could only 
accept justified six level tape. An optional reader was 
available to accept Justowriter tapes but as far as is 
known only one was ever sold. 


Input sources for second generation devices are: (1) direct keyboard entry, 
(2) keyboard-controlled perforated paper tape, and (3) computer-controlled 
paper or magnetic tape. Early systems used entry by keyboards directly 
interfaced with the machine and were designed for use by skilled operators 
who understood the function codes required to typeset copy and who could 
divide the copy into justified lines. This method was slow because of end-of- 
line justification decisions and extra coding strokes had to be made 
manually, (Photon 200 series). 


As faster, and more flexible phototypesetters became available, the use of 
keyboard perforating units for paper tape became more popular. These units 
changed keyboard text entry into an off-line operation and made the use of 
multiple keyboards for a single phototypesetter possible. This method also 
allowed the typesetter to operate at its maximum speed rather than that of 
the keyboard operator. 


24 Composition Input 








Photon’s 713 series of phototypesetters incorporated 
the necessary logic to produce hyphenless 
justification. With switches on the control unit, func¬ 
tion commands could be over-ridden in order to re¬ 
set copy without rekeyboarding. 


Perforated paper input tapes may also be generated by general purpose 
computers. These computers normally provide the flexibility and speed to 
perform composition work on a variety of large, complex typesetting jobs as 
well as other business applications. General purpose computers are capable 
of taking information from computer-generated data banks, such as 
telephone directories or parts catalogs, and, under the guidance of an 
appropriate program, prepare it for phototypesetting without the step of 
keyboard entry. Elimination of this step greatly enhances the efficiency and 
accuracy of the total system, as keyboard entry has historically been a 
bottleneck in the photocomposition process. 

An additional input source available with some phototypesetters is 
computer-generated magnetic tape. Magnetic tape carries the same coded 
input data as paper tape and permits much higher entry speeds with 
improved handling characteristics. 

Character selection consists of a light source, master character set, and 
control logic to synchronize the light source and character set. The light 
source is of high intensity and may be continuous or may operate 
stroboscopically with a flash duration of one or two microseconds 
(millionths of a second). The light source normally used is a xenon flash 
lamp or flash tube. For larger point sizes the lamp is flashed twice (multi- 
flash) to assure that the film is fully exposed for uniform character blackness 
(density). 


Composition Input 2 5 


A basic phototypesetting system: light source, lens, 
prism or mirror and photosensitive material. 


MASTER CHARACTER SET 


LIGHT SOURCE 


PRISM 


Master character sets are transparent negatives (clear characters on opaque 
backgrounds) which supply character images in the typeface or font desired. 
Master character sets are stored on: (1) rotating discs, (2) rotating drums 
with interchangeable film strips, (3) rotating turrets, or (4) stationary grids. 
In all cases, master character sets are interchangeable to permit insertion of 
desired fonts not already in the machine. Some phototypesetters offer 
multiple character arrays with up to five discs or four grids. 



OUTPUT MEDIA 


The process of character selection begins by coded information from the 
input source being read into the machine’s control logic. As the logic unit 
recognizes the character and font code, it determines where the desired font 
is located in relation to the light source. In machines with single character set 
arrays, all available fonts are located in front of the light source. Multiple 
character set arrays may find the desired font elsewhere and move it into 
position before the light source. Simultaneously, the logic unit is selecting 
the proper light source, when more than one is available, and positioning the 
apertures, mirrors, and lenses used to isolate the desired character. 


As the proper character image approaches the light source, most systems use 
a series of timing marks on the master character set or timing pulses to 
synchronize action between the light source and the character image. On 
some machines the timing mechanism will cause the character to be 
projected to stop in front of the light source. However, on most machines the 
extremely fast stroboscopic light source is sufficient to obtain high quality 
output by simply flashing at the exact moment the character image is in 
front of the light source. Once the proper character is selected and its image 
is optically initiated, image output control begins. 


Image output controls and positions the character image on the output 
media. Components of the subsystem include an optical projection 
mechanism to focus, direct, and magnify the projected light rays and a 
spacing mechanism to accurately provide intercharacter and interword 
spacing. 


26 Composition Input 






i. 

_ J 

Ll _ 

1 L 

» _ 


J 1 _ 

L J 

1 1 

_l -L 

1 

- J - 

1 L « 

X J 

L l. 

— 

J Li 

- - 1 

1 _ 

- J 


J_1 

.1 X. 

LJ 

L L 


J _ .1 

LlL 

1 -J 

' X 

J 

L L L 

1 X L 

L I 

X « 


L_1 » X 

1_ 

_» 

J 

LlL 

JL _J. 

Ll 1 


J_ 

JL -1 

_» 

*-1 


J _i J 

J L_!_ 

i J 

L 

-1 

J J JL 

L 1_l~ 

L - 

1 J 

_L 

_i L 

*-1.1 

J_‘ 


GRID 


Optical projection mechanisms vary significantly among phototypesetters. 
Rotating character sets normally carry several fonts around the rotating 
surface. The stroboscopic light flash illuminates a complete row of 
characters rather than isolating individual characters. A series of apertures, 
mirrors, or lenses moves into a unique set of positions in response to function 
codes on the input tape. The desired character image is thus guided into the 
optical path while all others are deflected away. As the number of fonts on 
an individual rotating character set increases so does the complexity of the 
optical system required to isolate and project individual characters. 

Stationary grid systems are the most complicated systems of all for optical 
character isolation. Since character images on the grids are stationary and 
not rotating past the light source, a continuous light source is used. Light is 
distributed evenly over the grid by condenser lenses specifically designed for 
this purpose. The effect is illumination of an entire matrix of character 
images rather than a single row. The light rays from each character are 
formed into a beam of virtually parallel rays by a series of lenses called a 
collimator assembly. The rays then enter a series of eight pairs of optical 
wedges designed to isolate the desired character image. Directed by codes on 
the input tape, the wedges move to a unique set of positions for that 
character. The refractive angles on the wedges guide light rays from the 
appropriate character along the optical path and deflect all others away. A 
decollimator assembly is used to converge the parallel light rays and 
restructure the character on the output media (Linofilm Quick series). 

In typesetting, point size is a measure of type height. It measures the 
maximum height a font requires when adding together the maximum 
distance reached by characters extending both above and below the base line. 
A 72-point font requires one full inch to accommodate all characters within 
it. Normal text type ranges from 5 to 12 points and most newspaper and 
book headings are less than 18 points. Characters in point sizes over 18 
points are normally considered as display characters. 



Photographic matrices come in varying formats: disc, 

drum, film strip, turret and grid. Font grids installed on the Mergenthaler V-I-P. 


Composition Input 27 











All phototypesetters have available a range of character point sizes, called 
the point size range, but differ significantly in the number of specific settings 
available and the method for changing point size. Systems vary from 
manual insertion to highly automatic selection of numerous settings. 

Systems that require manual changing of point size permit by far the most 
straightforward and inexpensive design of the image output control 
subsystem. In most of the low-priced text-oriented machines no mechanism 
to control point size is required, as manual changing of the master character 
set and the set-width gears is required. In several machines, manual 
adjustment of point size is accomplished by the setting of levers to 
correspond to the desired point size. Movement of the levers causes lenses in 
the main optical system to change position and vary magnification of the 
output. 

In the Fototronic system, variations in magnification are accomplished 
automatically by movement of two lenses. Directed by function codes on the 
input tape, the movable lenses are driven along the optical axis by motors 
until positioned in the appropriatelocations for the desired magnification. 



The Harris-Intertype Fototronic. Input to this high 
quality phototypesetter is via special counting 
keyboard or computer. 


The most commonly used method of obtaining different point sizes is a 
multi-lens turret. The turret normally holds eight or more lenses which 
provide a range of magnification levels. Each magnification level relates to a 
specific point size. The approprite lens is rotated into position in response to 
a coded command from the input tape. Lens turrets provide good 
typographic flexibility with reasonable photomechanical simplicity. 


28 Composition Input 







There are numerous methods by which proper intercharacter or interword 
spacing can be achieved. Two basic components are involved, the 
mechanism that physically sets and advances character images across the 
output media and the logic system which controls the amount of advance 
and synchronizes it with the stroboscopic light source. 

A lens carriage which moves laterally across the output media to locate the 
character properly along the typeset line is the most inexpensive method 
available. After each character is set, the lens carriage advances a distance 
equal to the appropriate character width and is ready for the next character 
image. Master character widths are either permanently hardwired into the 
circuitry of the phototypesetter or are available through changeable width 
circuits called width plugs. The character widths are then applied to the 
carriage advance through a pair of set gears that must be changed for each 
change in point size. 

Another simple method for placing characters along the line is a positioning 
mirror that rotates to deflect the image the desired amount. Character focus 
is maintained by a focusing lens that moves in synchronization with the 
mirror. Rotation of the positioning mirror is controlled by a second optical 
system that generates electrical pulses to indicate the amount of rotation 
occurring. Rotation will stop when image deflection equals the appropriate 
character width. Perforated program tapes are used to load character widths 
into the machine memory in coded form. A new character width tape must 
be read into memory for each change in point size. 


A third technique for correctly positioning characters along the line of copy 
involves a carriage that advances the light source and master character set 
literally in relation to the output media. The carriage advance is determined 
by the character width contained in the device’s memory. Loading of the 
memory from special coded tapes before typesetting begins is again used. 







After the 200B, what...tape of course. Photon brought 
out additional units that only ran from tape. The 513 
and 560, shown, were part of a series of units. The 
560 was the first computer slave. 


Immediately adjacent to the master character set are two movable apertures, 
one with two horizontal slits and one with a single vertical slit. Light passes 
through these apertures only at the intersection of the slits. Directed by the 
machine’s control logic, the apertures are positioned so that light rays from 
only the desired character image pass through the intersection. However, 
due to control circuitry built into the machine, the projection of a specific 
character’s image need not originate in precisely the same location each time 
the character is used. Timing of the flash and movement of the apertures can 
be controlled to vary the optical path between the displacement limits. 
Character width, interword spacing, and intercharacter spacing are all 
considered during movement of the optical path. Only after maximum 
displacement of the optical path has occurred will the carriage advance. 

Interline separation or leading is simply the distance separating sequential 
lines of typeset copy. After each line is set, the end-of-line function code 
from the input source triggers a mechanism that advances the film or paper 
the proper distance. Leading information is often entered through a manual 
setting on inexpensive phototypesetters but is generally entered as a function 
code on larger, more versatile machines. 

Third generation CRT phototypesetters produce typeset copy by means of 
electronic selection and exposure of character images on the face of a CRT. 
The major subsystems comprising a CRT phototypesetter include date 
input, character storage, character generation, and output pagination. 

CRT phototypesetters usually operate in an off-line mode, with computer¬ 
generated magnetic tape as their input source. The tape contains the same 
character, function, and interline spacing codes associated with second 
generation devices. Magnetic tape is also used for font loading with those 
CRT phototypesetters using digital character storage. 


30 Composition Input 










































The Linotron 505 has an optional capability for magnetic tape input, but is 
primarily designed to operate from perforated paper tape input. In many 
ways, the Linotron 505 is more closely related to second generation 
phototypesetters than to other CRT devices. 

Character storage in third generation phototypesetters is either 
photographic or digital. Photographic storage requires the positioning of 
master character sets in the optical path. Access to fonts not already stored 
within the machine requires manual changing. 

Digital storage defines character shapes by digital coding read and processed 
by a computer. The information required to describe characters digitally is 
so large that peripheral magnetic disc systems are required to maintain a 
reasonable font selection within control of the computer typesetting system. 
Magnetic disc systems are currently used in two ways. (1) directly interfaced 
with the main computer system or (2) directly interfaced with the 
phototypesetter. 

Magnetic disc systems peripheral to the computer enter digital font 
information into the phototypesetter by magnetic input tape. Upon entry, 
the information is stored in the magnetic core of a small control computer. 
Here the information is available on-line as required. 

Generation of character images for CRT phototypesetters occurs from 
electronic projection of character descriptions stored within the machine. 
Phototypesetters that use photgraphic storage employ two CRT’s to 
generate character images. One CRT is a scanning device that describes the 
character and the second displays the character for exposure of the output 
media. Machines that store characters digitally use a single display CRT. 


Composition Input 31 





































































































PENTA PRISM 


POSITIVE FIELD LENS 
EYE LENS 



OUTPUT MEDIA 



REFLECTION 

MIRROR 


PENTA 

PRISM 


DECOLLIMATOR 

ASSY 


OPTICAL 
WEDGE ASSY 


COLLIMATOR 
LENS ASSY 


CHARACTER GRID 



CONDENSER 


ILLUMINATION ASSY 


The Linofilm Quick imaging system uses a stationary 
matrix. 


The Super Quick and Wide Range Super Quick were, 
and still are, expanded versions of the basic Quick 
photosetter. The unit on the left is the optional Tab- 
matic unit for tabular composition. 



■■H 


;•.. 


32 Composition Input 




















Also part of character generation for CRT phototypesetters is the system 
used to change point sizes. In scanning systems, height adjustments are 
made by changing the length of the stroke. Width changes are more complex 
and must be accomplished by either, (1) increasing the dimensions of the 
electron beam while spacing the strokes further apart, or (2) maintaining a 
constant electron beam size and using more strokes to describe the 
character. 



CRT phototypesetters designed for photographic character storage vary 
point size by using a constant stroke width in the display CRT and 
increasing or decreasing the number of strokes used to describe the 
character. This is done by varying spacing of the vertical scan lines in the 
scanning CRT in relation to the point size. For example, for a 5 point width, 
the master character image would be scanned in half as many strokes as used 
for a 10 point width. When combined with a constant beam width in the 
display CRT, the result is a character width or point size in proportion to the 
number of scanning strokes. 

Output pagination refers to the method used to generate a full "page of 
typeset copy. The primary methods used are: (1) stroke, (2) line, and (3) 
page. 

In operation, each stroke displayed on the face of the CRT is located in 
exactly he same position. No horizontal movement of the electron beam 
occurs at any time. A mechanically driven prism lays down each vertical 
slice of the character one at a time. A series of timing lines synchronizes 
movement of the prism with generation of each stroke. 


Composition Input 3 3 






INDEX TUBE 


The Mergenthaler 1010 imaging system. 



The imaging system of the Linotron 505. Note that it 
is a hybrid: using a matrix to tell the CRT how a 
character should be generated. 



The line method utilizes horizontal deflection of the electron beam to typeset 
a complete line and a mechanical system to advance the film or paper 
between lines. This method is much faster than the stroke method because 
electronic beam deflection is faster than mechanical action. However, the 
extra speed is achieved at the cost of a more expensive, larger CRT and more 
complicated circuitry to compensate for distortion of the electron beam as it 
moves away from the center of the tube. 

The page method uses both horizontal and vertical deflection of the electron 
beam to set a complete area or page of type without any mechanical 
movement. The output film or paper is advanced only between pages. While 
the fastest typesetting method of all, it is also the most expensive. 

The Fototronic CRT uses a combination of the line and page methods. It 
combines beam motion with motion of the output film or paper for both 
horizontal and vertical placement of characters. Horizontal character 
placement is accomplished by beam deflection across the full 11.5-inch width 
of the CRT display tube. In addition, the output film or paper is on a 
carriage that can be shifted sideways if a wider line is required. 

There are several developments in technology and applications that will 
significantly affect the phototypesetting industry, although it is too early to 
predict the nature or extent of their impact. 


34 Composition Input 











3. Keyboard to tape systems 


To understand the wide variety of keyboards available today, and to 
appreciate some of the trends toward new equipment concepts, it is useful to 
have some background on the general development of keyboard systems as 
used in the typesetting environment. For our purposes here keyboards, 
keyboard systems, keyboard machines and keyboard devices are defined as 
those pieces of machinery which create a machine readable record as their 
primary output. This record is then subsequently processed by some other 
device or system. Keyboard elements are those subassemblies consisting of 
at least 64 keys mounted on a frame. 

Initially, virtually all typesetting was done on hot metal linecasters made by 
Harris-Intertype and Merganthaler. The operator used a keyboard 
component that was an integral part of the linecaster, and as information 
was keyed, a line of brass type mats was assembled to form a line from which 
a lead slug was cast. These lead slugs were then locked in a chase and this 
formed a complete galley of information ready to be printed. The maximum 
line length was limited by the capacity of the linecaster and was usually 30 
picas. To achieve uniformity of line, it was necessary for the operator to key 
sufficient charaters to fill the line (supply enough characters to form a 
complete slug). On the other hand, if too many characters were keyed, the 
machine stopped due to an overset condition. This “enough but not too 
many” measure (called the hot zone) is simply a method of telling the 
operator when enough characters have been keyed to form a line. Once the 
hot zone was reached, it was up to the operator to determine how to end the 
line-to either finish the word, or to hyphenate. 

Three major events moved typesetting away from the use of manual systems 
to the use of modern phototypesetting devices of today: 

1) The first significant event to influence typesetting procedure was the 
demonstration that linecaster could be adapted to operate from punched 
paper tape readers. This permitted the tape to be made elsewhere on 
relatively inexpensive keyboard systems and then processed by the 
linecaster. Three major advantages were: 

a) the tape could be prepared by people with less skill than a linecaster 
operator 

b) the keying speed was faster when only punching tape 

c) the linecaster could operate at its maximum speed, hour after hour. 


Composition Input 3 5 



The use of independent keyboard systems operating a tape driven linecaster 
significantly increased the effective “throughput” (or characters per dollar) 
of the system. 

Several manufacturers offered special keyboard systems to serve this 
growing market. The first such devices were blind counting keyboards which 
produced justified tape. In use, both the length of the line to be set and the 
size of the type to be used was determined and the keyboard device adjusted 
for these parameters. As the operator keyed material into tape, an internal 
counting mechanism kept track of how much of the line had been consumed. 
When the hot zone had been reached, an indicator informed the operator 
that it was time to end the line, and the operator had to either finish the word 
or hyphenate, and insert a line end code. These blind keyboard systems 
consisted of a keyboard element mechanically connected to a paper tape 
punch, with a counting mechanism included. The operator had no visibility 
of the material being keyed, except to read and translate the code holes 
punched in the tape. In an attempt to furnish the operator with visibility of 
keyed material, some manufacturers connected tape punches and counting 
mechanisms to typewriters, and offered hard copy counting keyboards. 
These hard copy systems were more expensive, and in general they found 
their best application in training new keyboard operators or in preparing 
complex material. 

2) The second event concerned the use of computers, which mostly 
affected high volume typesetting, such as newspapers. It was found that 
operators could key material up to 50 per cent faster if they did not have to 
concern themselves with “line end decisions”. By switching off the counting 
devices in the keyboards, the operator could key non-justified tape. This 
non-justified tape was then fed into a special purpose computer, which had 
been programmed to assemble the characters into a specific line length. The 
computer also contained a set of logic rules to govern hyphenation. The 
computer/then read the non-justified tape, examined each character to 
determine how much space in the line it would consume, hyphenated words if 
necessary, and produced a justified tape to operate the linecaster. 
Compugraphic was the pioneer in the field of computerized typesetting, and 
the Justape and Justape Jr. led the field in competition with IBM. The 
application of the computer to the typesetting environment permitted an 
important increase in keying speed, resulting in an effective lower keying 
cost. Manufacturers offered both blind and hard copy keyboard systems to 
serve this market application, although in many cases these machines were 
simply counting keyboards with the counting mechanisms removed. 

These two events greatly increased the effective performance of hot metal 
typesetting. The first event (independent tape controlled operation) removed 
the human limitation and allowed the linecaster to operate at its maximum 
speed; the second event (computerized processing of non-justified tape) 
eliminated the concern for line end decisions and allowed the human to key 
at appreciably faster rates. For comparison, a good linecaster operator could 
key at approximately 6,000 characters per hour, while a good keyboard 
operator using a non-counting keyboard could reach rates of 18,000 
characters per hour. Thus the use of non-counting keyboards combined with 
a typesetting computer achieved a three times increase in typesetting 


36 Composition Input 



throughout. In addition, the requirement for highly skilled operators was 
reduced, which therefore lowered the cost of keyboarding. 

3) The third event to consider was a basic change in printing technology and 
the resultant impact on typesetting. Hot metal linecasters prepared lines of 
type, later assembled into galleys, which were used with letterpress or 
sterotype rotary printing equipment. 

However, raised type was not usable in an offset press, so an intermediate 
step was required between the assembled typeset galley and the offset plate. 
Most users simply ran a “repro copy” from the raised type, photographed it, 
and used this negative to make the offset plate. 

Phototypesetting was born with the development of machines using photo 
flash techniques to record the selected character image on a piece of film or 
paper. This output film or paper then became the equivalent of the typeset 
galley. The character images were selected from font discs or grids through 
which the light was flashed and the size of the character image could be 
altered by choosing lenses of different sizes, which finally focused the image 
on the film. Initial phototypesetters were functional replacements of the hot 
metal machines and, therefore, the preparation of input tape was the same 
for either hot metal or phototypesetters. 

Subsequent phototypesetters offered far more flexibility with respect to font 
size and the number of fonts available. This increased the complexity of 
keyboarding, and has given rise to more sophisticated keyboard systems to 
produce this tape. The role of the computer has also expanded, and in 
addition to performing the simple hyphenation-justification tasks, the 
computer can now manage many of the phototypesetter functions. 

In general, the more straightforward phototypesetters manufactured today 
have the hyphenation-justification computer included, and consequently 
virtually any kind of tape is acceptable. These machines are mostly used in 
straight matter text environments, such as the production of newspapers. 
The more sophisticated phototypesetters, which require the use of justified 
tape, are usually matched with the complex counting keyboards, or with 
computer systems. 

Today, a bewildering selection of equipment is available. The user must 
choose between counting or non-counting keyboards, a host of typesetting 
computers, and typesetters that accept both justified and non-justified tape. 
New keyboard systems and computer systems introduce the question of 
recording media — magnetic tape, punch tape, printed tape, punch cards — 
are all available and they all have their place. As the number of options and 
permutations proliferate, it becomes increasingly important to understand 
the entire system, and its application. This “system appreciation” is 
imperative before an intelligent recommendation can be made relative to the 
phototypesetter and its attendant keyboard devices. 

Recording Techniques 

All recording techniques in use today are based on the absence or presence of 
a recording “bit”. A bit, by definition, is a Binary Digit, and binary digits 


Composition Input 37 



are the fundamental units of a numbering system which uses 2 as a Radix. 
This base 2 system (using the “0” and “1” as the digits) is particularly 
applicable to recording schemes, since these two digits can be stored or 
recorded by a variety of mechanical or electronic devices. For example, relay 
contacts can be open or closed, a pulse can be absent or present, a magnetic 
field can be polarized north or south, a light can be off or on, etc. 

Different numerals, alphabetic characters, or symbols can be recorded by 
assembling these bits in various combinations, called codes. The number of 
different codes available in any system is a function of the number of levels 
used - a level corresponding to a significant place in the system. For 
example, a code system using 5 levels has a maximum number of 32 code 
combinations. A code system using 6 significant places or levels offers 64 
possible combinations, and so on. The number of possible code 
combinations is limited only by the number of levels or significant places 
used. 

Since our concern is with typesetting, our attention will focus on the code 
combinations most commonly used by typesetting equipment. The original 
code system was developed by the Teletype Corporation, and is referred to 
as a TeleTypeSetter, or TTS code. Six levels are available, so the system 
offers 64 codes. These codes are punched in paper tape, with a hole being 
equal to “1” and a no-hole being equal to “0”. 

TTS tape is seven-eighths inches wide, and each of the data holes or bits is 
.072 inches in diameter. The sprocket hole is .046 inches in diameter, and 
while the primary use of this sprocket hole is to mechanically move the tape 
through the punch or reader, it also serves as a reference or timing bit to 
identify a character position in the tape. Character density is 10 characters 
per inch. Tracks are numbered 0 thru 5 (some manufacturers number the 
tracks 1 thru 6) and the sproket hole is in the center of the tape. All the bits 
that constitute a character are punched at the same time, along with the 
sprocket hole. 

As mentioned above, a 6 level TTS code set offers 64 unique code 
combinations. Since a typesetter may contain over 100 characters in any 
font, plus 10 to 20 control codes, it is necessary to find some way to expand 
the number of useful codes available. This is done by using a “precedence” 
technique. When a precedence code is punched in the tape, it defines the 
meaning of all subsequent codes. The TTS system uses “shift” and “unshift” 
as precedence codes. While this does not change the number of unique 
combinations in the system, it does increase the number of useful codes, 
since any given bit combination can now have two meanings. For example, 
in the shift case, a hole in track 0, 1 is the numeric character “ 3 /s”. In the 
unshift case the same hole combination is the numeric character “3”. Of the 
64 codes in the TTS system, 2 are dedicated as precedence codes, one is the 
rub-out code and one is the tape feed code, resulting in 60 codes in the shift 
case and 60 in the unshift case, or 120 in all. Some codes, like elevate, return, 
thin space, etc., are the same in either the shift or unshift condition. 

Some systems use a third precedence code (sometimes called red ribbon shift 
or control case shift) which allows 59 codes in each case, or 177 useful codes. 


38 Composition Input 



Recording Media 



Paper tape is one of the original recording media. The paper itself is usually 
furnished in 1000 foot rolls, and is normally 2 Vi to 4 mils in thickness. Paper 
tape is available in standard widths of eleven-sixteenths inches, seven-eights 
inches and 1 inch. While the actual recording format is standard the world 
over (data holes .072 inches, sprocket holes .046 inches, holes on one-tenth 
inches centers) there are a number of methods used to number the tracks. In 
addition, some schemes use the sprocket hole on a line tangent with the date 
holes. The scheme used will vary with industry and application, and virtually 
any tape width will be found in any industry. In general, the primary 
applications of paper tape are: 


o 




2 

Feed 

3 



Data holes or bits 


..07 


Sprocket 


2 Character Or Row 


eleven-sixteenths inches wide 
seven-eighths inches wide 
seven-eights inches wide 
one inch wide 


5 level communications 

6 level typesetting 

7 level data processing 

8 level data processing 


Paper tape is widely used due to its low cost, coupled with the modest cost of 
punches and readers. A 1000 foot roll of tape costs approximately $1.00 and 
will contain 120,000 characters. The cost of the punches and readers will 
vary with the operating characteristics and quality, with punches of the 15-30 
characters per second vaiety costing $350-600 and ranging up to $2500 to 
$9000 for punches operating at speeds of 150 to 300 characters per second. 
Readers range from $250 to $500 for speeds of 15-30 characters per second. 
In general, paper tape finds its most popular application in those areas where 
the recording and reading rate is not high, and low cost is an important 
factor. 


MagneticTape 


This concept of recording and retrieving information was pioneered by the 
data processing industry, and was designed to satisfy the need for fast input 
of data to a computer. The tape is usually furnished in lengths of 2400 feet 
(wound on a 10 inch reel) and is Vi inches wide by 1 Vi mils thick. A reel of 
tape costs around $35, but the information on the tape can be erased and the 
tape reused many times. As with paper tape, the characters are made up of a 
pattern of bits. The bits are represented by reversals in magnetic flux 
direction. This is done by moving the tape past an electromagnet (called a 
write head) and reversing the flow of current at the time a bit is to be 
recorded. Heads are arranged across the tape, with one head for each track 
or level. Information is recorded by the write head, and is retrieved by the 
read head, both of which are contained in the tape deck. Magnetic tape 
permits a far higher read-write rate as well as a higher character density than 
does paper tape. Today, the most common kinds of tape formats are: 

7 track 200, 556, or 800 characters per inch (CPI) 

9 track 800 CPI 

9 track 1600 CPI (requires special read-write techniques) 


Composition Input 39 



The tape is moved past the read-write head at a constant speed during the 
writing or reading operation. The tape stops in the gaps between blocks. As 
the tape is read, a flux change in any of the tracks (or any combination of 
tracks) signifies the presence of a character. This makes the tape “self 
clocking” and eliminates the need for sprocket bits or timing bits to be 
recorded along with the character. The speed at which the characters can be 
recorded or retrieved (transfer rate) is a function of the character density 
(characters per inch or CPI) and the speed of the tape (inches per second or 
IPS). In a system using a tape speed of 40 inches per second, and a character 
density of 800 char per inch, the effective transfer rate is 40 IPS X 800 CPI 
or 32,000 characters per second, either read or write. Tape drives are 
currently available with speeds from 7,000 characters per second up to 
180,000 characters per second. Mag tape drives are usually furnished with 
both read and write electronics. Depending upon the characteristics of the 
unit, a mag tape drive may cost anywhere from $4,000 to $40,000. These 
units usually find application in those environments where large amounts of 
data must be recorded or retrieved at high speeds, and where prime cost is 
not a significant factor, due to large data volume. 

New lower cost tape drives (called incremental units) are now being offered 
to bridge the gap between the slow rates of paper tape devices and the 
extremely fast speeds of the magnetic tape drives discussed in above. These 
units record information on magnetic tape in a computer compatible format, 
however, in place of recording a block of data at once, they record a single 
character at a time. The recording can be done at any random rate up to the 
maximum, and the tape increments (moves) one character position for each 
character recorded, thus the name incremental. Generally they accept data 
at rates up to 300 characters per second, and the tape can then be loaded on a 
computer and read at maximum computer speeds. Some incremental 
magnetic tape units have both read and write functions, and typically ’’slow- 
read” the tape at rates up to 6,000 char per sec. These units sell for around 
$5,000 to $10,000, depending upon their characteristics, and are generally 
suited for applications where the information is recorded at fairly low speeds 
and read into a computer at high speed. 

Another approach to the problem of low cost recording coupled with fairly 
fast reading speeds is based on the use of !4 inch magnetic tape. This tape is 
similar to that used in home entertainment systems, with an important 
difference being the characteristic quality of the oxide coating. As before, 
characters are made up of groups of bits, and (like Vi inch mag tape) the bits 
are recorded by changing the flux direction as the tape moves past the read 


GaP D □□□□□□ □ Gap □ □ □ Etc - 1/4" 

Bits 12345678 123 i 


A representation of seven track magnetic tape. 


40 Composition Input 



head. Bit patterns are recorded sequentially in tracks. The tape pattern 
provides a start-stop gap between characters, so that the system may record 
and read information a character at a time. Generally these systems are all 
based on an 8 level code, thus 8 bits are always in the string. The system is 
self clocking, so no sprocket or timing bits are recorded. The recording 
density can be as high as 800 bits per inch, but since the characters are 
sparated by gaps, the effective density is around 40 characters per inch. 
These units offer recording rates as high as 1600 char per sec. The price of a 
combined read-write unit is approximately $3,000. 

Magnetic cassettes also evolved out of the home entertainment field, 
and were pioneered by Philips-Norelco. They are extremely easy to handle, 
and are very small in size considering the amount of information they can 
store. In principle, they are identical to the !4 inch mag tape machines in 
that they record the bits in serial strings, with gaps between characters. 
Cassettes use tape only Vs inches wide. They can store up to 80,000 
characters in a single cassette, and the tape is re-usable many times. The cost 
of a single cassette varies from $3 - $12. Cassettes are widely touted as the 
replacement devices for punched paper tape, and there are examples where 
the cost of the recording and reading units appear competitive with punched 
tape. A unit combining both read and write functions (such as the Sykes 
unit) may be priced as low as $750, and offers write speeds of up 500 char per 
sec. Their primary disadvantage seems to be the lack of an industry wide 
standard governing the record-format. Various manufacturers of cassette 
equipment may use different track-bit-code allocations, which means that 
cassettes may not be interchangable between devices made by different 
companies. 

While the term cassette (by popular definition) means the Phillips-Norellco 
style, more and more companies are offering their own unique versions. 
IBM uses a special cassette on their MT-ST-MT-SC equipment, where the 
tape is Vs inches wide and is moved by means of sprocket holes punched in 
the edge of the tape,similar to motion picture film. Characters are recorded 
in parallel across the tape, similar to punched paper tape. Invac offers a 
special cassette using l A inch magnetic tape, where the tape moves from reel 
to reel, and the characters are recorded in bit sequential fashion. 

Keyboard Primer 

This section deals with keyboard systems that are now in common usage. 
Some of these keyboards have been available for many years, while others 
are relatively new. It is impossible to compare every feature of every 
keyboard to its competitor, so this section is intended to give an overview of 
keyboards currently in use. 

Keyboard element — For hard copy devices, this means the kind of 
typewriter used. For blind devices, this refers to the keyboard component. 

Secretary shift — This defines the shift and unshift configurations of the 
keyboard element. Some manufacturers use a separate key for both shift and 
unshift commands with the attendant code generated as either key is 
depressed. Secretary shift commonly means that the shift code is generated 
when the key is depressed, the unshift code when the key is released. 


Composition Input 41 



TAPE IDENTIFICATION CHART 


PUNCHING 

DIES 


CHANNEL 

NUMBERS 


CHANNEL — 

FRAME**- 

ADVANCED _ 

FEED HOLE 

RUBOUTS < 


TAPE FEED < 

ADD 7 — 
ADD 8 — 


TAPE PUNCH 

-►OOOOO o ooo 



< 


8 


128 


64 


0 

6 


1 


32 




16 


8 


F 

E 

E 

D 




1 


) 




ONE INCH TAPE 


FLOW 
OF TAPE 


t 


FEED HOLE .4375" 
FROM GUIDE EDGE 

GUIDE EDGE 


BINARY VALUES 


*A BIT IS EITHER THE PRESENCE OR ABSENCE OF A HOLE 
**A CODE MAY BE ONE OR MORE FRAMES 























SW>C£ H 8f>A«£ 1 SPACC 

~fis m mvt m -My 


zma § ucao ■ te*»: 
ueao 1 ipr 1 mm 


fan at 


plash 1 mat # wi 
onu 1 just I; mm 


mm 

tm 




ill 








PRINT 







UM 



: WHiai r 



■ ' i i 

: TA 

1 

m 

• 

# 

n§ 

- m 

III 4-i 

I 

4 k 

1 


Datek of England manufactures an extremely ex¬ 
tensive line of input keyboards and peripheral equip¬ 
ment. One peripheral that is of increasing interest is 
the line printer. This is the Datek version of a device 
to accept input tape and print out its contents for 
proofreading purposes. 


Additional keys on the Dual Image keyboard permit 
single stroke command of all Compugraphic 4961TL 
typographic functions. Depression of one key often 
produces two printed codes. 





mmvx 


we* cm,c; 


imz izmm 


m m 


*: c a : 


cm# um. 




A closeup of the key set of the Mergenthaler V-I-P 
keyboard. Note the extra rows of character keys. The 
indicator in the center displays line length for 
justification purposes. This is counting keyboard. 


Composition Input 43 


















Back space — This feature allows the tape to be moved back one character 
at a time from the keyboard element. Most keyboard systems have some 
provision for reversing the tape, but not all do this by depressing a key. 

Repeat key — This key, when depressed with certain other keys (or in some 
cases, any other key) causes that particular key to repeat until released. 

Format storage — This is the ability to store, either on tape or 
electronically, a number of predetermined character sequences for changing 
line length, leading, font size, etc. 

Multiple codes - one key — This is the ability to generate a 2 or 3 character 
sequence from a single key on the keyboard element. These are usually pre¬ 
wired at the factory and relate to some specific typesetter function, for 
example ‘flash only’. 

Widths mixed from the keyboard — Some keyboards can store different 
character width values (usually in the form of width plugs) and the operator 
can select the one desired by merely pressing a key. Width plugs usually 
relate to a particular font. 

Code format — This refers to the number of channels or holes in 
punchepaper tape. Where an entry appears it means that the unit is available 
in either 6 level TTS code, 7 level Friden, or 8 level ASCII. 

Hot zone indicator — The method used to tell the operator when the line end 
decision should be made. 

Maximum line length picas — While true line measure varies with the 
counting technique used, most counting keyboards are classified by the 
maximum line length that can be counted, as measured in picas. 

Type of display — In some cases, a typewriter; in others a moving display of 
the last 16 characters. 

Number of characters — Refers to the number of different symbols that can 
be displayed to the operator, and has nothing to do with the number of 
characters in any given font set. 

Notes on counting 

To ensure that the finished typeset galley is well balanced and asthetically 
pleasing, it is necessary to use a system of differential character widths. In 
this way the larger characters, such as the W and M, occupy more horizontal 
space than the smaller characters, such as the I and J. Two methods are in 
use today. 

Unit cut fonts 

Sometimes called standard, or standard cut mats. In this system, the widest 
character in the font (the upper case ‘M’, also called the ‘em quad’) is divided 
into 18 equal parts, and multiples of the one eighteenth dimension become 
the width of the different characters in the font. 


44 Composition Input 



Non-unit cut fonts 



One version of the Datek tape perforating keyboard 
(marketed in the United States by VGC). The 
keyboard is noiseless and the punch is covered. 
Operators were raising the cover in order to hear 
some indication that a key registered. Manufacturers 
have found that this “sound of accomplishment” is 
quite important. 


Sometimes called multiface mats, this system is based on the division of the 
em quad into 32 units. Multiples of the one-thirty second dimension become 
the width of the different characters in the font. For example, the lower case 
T may be 7 units (seven-thirty seconds of the em) with the upper case ‘M’ 
being a full 32 units. This ratio of character widths is true only for a given 
font size. That is, the T may be 7 units in 12 point type, but only 5 units in a 
16 point font. The multiface system allows greater precision in typesetting, 
since the width relationship of the characters varies with the point size of the 
font. 

Blind, counting keyboards 

Fairchild was the first company to offer keyboard systems to prepare the 
tape. The keyboards were designed to match a given linecaster, and were 
tailored to the kind of font being used, (unit cut, or multiface). Where the 
keyboard was used with a unit cut standard font, the keyboard was simply 
adjusted for the length of the line and the size of the type. Where the keboard 
was used with a multiface font, a ‘width plug’ was inserted to program the 
keyboard for the character width relationships, and the keyboard was 
adjusted for the length of the line to be set. Each font had an appropriate 
width plug. These machines were intended for the production environment 
and very little flexibility was built into them. With multiface fonts, it was 
necessary to manually change width plugs when changing font sizes. 

Subsequent keyboards offered more flexibility, and multiface machines were 
now offered with the capability of having more than one width plug inserted 
at once. The Fairchild Universal 210 could hold two, and the Merganthaler 
Lino-Quick could hold 4. This permitted widths to be mixed from the 
keyboard, since the desired width plug could be selected by simply pressing a 
key. Blind keyboards offer no visibility of the data being keyed, although 
some operators are proficient at reading the holes in the tape. Most 
keyboards furnish indicator lights to display the status of the shift condition, 
upper or lower rail, etc. Line end decisions are made by the operator once 
the hot zone is reached. Most of these keyboards can be used in the non¬ 
counting mode. 

Notes on hand copy 

There are a number of different viewpoints on the subject of hard copy, and 
the opinions are varied and somewhat subjective. In essence, hard copy 
offers the keyboard operator visibility of the material being keyed. Most 
hard copy devices use a typewriter, which means that a typed page is 
prepared which corresponds to the information being recorded. Typewriters 
can display upper case, lower case, numbers and symbols. The typewriter 
character set is limited to the number of character positions available. The 
IBM Selectric machine furnishes 88 positions, while some of the type bar 
machines offer as many as 96. Since the printing is done by one mechanism, 
and the recording by another, it is impossible to guarantee that the 
information being typed is exactly the same as the information being 
recorded. Because of the character set limitations it is not feasible for a 
typewriter to display control codes, such as upper rail, quad left, etc. 


Composition Input 45 



Proponents of the hard copy machines feel that the primary advantage is 
accuracy. Since the operator can see every character, it is very easy to 
correct the most common keying problem, the single keystroke error. The 
hard copy machines are very useful for training new operators, and in 
addition, it is easier to compose complex copy (like display advertising) 
when the information is visible. The opponents of hard copy maintain that 
visibility of the data actually slows down the operator, and faster sustained 
keying rates are possible if the operator can not see what is being keyed. In 
addition, the necessity of returning the carriage breaks the operator rhythm 
and reduces keying rates. (In counting keyboards, where the operator must 
observe a hot zone indicator and make line end decisions, this argument is 
hard to support. It’s extremely valid with non-counting keyboards). 

The trade-off seems to be between higher production speeds (blind 
keyboards) and increased accuracy (hard copy systems). The final 
configuration is determined by the kind of task. 

Hard copy, counting keyboards 

These devices were originally offered to furnish more accurate keying of 
tapes to drive hot metal machines. However, the advent of phototypesetters 
began to change the keyboarding problem, due to the greater flexibility of 
the phototypesetting systems. It was now possible to set both straight matter 
and ad copy on phototypesetters, however the keying task to accomplish this 
became much more complex. For example, it required only a single 
keystroke (Upper Rail) to go from normal face to bold face, since the line 
length, leading, and character width criteria remained the same. However, to 
go from 8 point to 16 point type could take as many as 20 keystrokes in a 
precise sequence, because of the line measure implication. Thus, keyboards 
were offered to address the problems of doing in-line width mixing, and the 
changing of character fonts and-or line length. These keyboards contain 
subroutines that are adjustable to match a number of different typesetting or 
composing tasks. 

The same counting procedures are used by these devices as with the blind 
counting keboards. Line end decisions are made by the operator once the hot 
zone is reached, and lights are used to display the status of the shift, upper or 
lower rails, etc. Most units can be used to produce non-counted tape. 

Blind, non-counting keyboards 

The advent of the typesetting computer created the market for these 
keyboards. The primary intention was for the keying of straight matter, 
where the operator made no line end decisions and speed was of primary 
importance. These keyboards permit the fastest posssible keying rate, and 
are limited only by the speed of the operator. While there is no visibility of 
the material being keyed, experienced operators seldom find this a problem, 
although new operators do. These are the simplest keyboard systems 
available, consequently they are the lowest priced. Some of these keyboards 
can be connected to slave typewriters to furnish hard copy, but this is not 
common in a production environment. Status lights are furnished to display 
the shift condition, upper or lower rails, etc. 


46 Composition Input 



A second type of blind non-counting keyboard was offered to prepare tapes 
for the more sophisticated phototypesetters. Rather than build format 
storage routines into the keyboards, the complex routines were stored in the 
computer used with the phototypesetter. In this way, four or five codes 
punched in the non-justified tape could cause the computer to insert the 
proper sequence of commands in the justified output tape. Keyboards with 
this feature usually have the function keys on a separate auxiliary keyboard 
and are sometimes referred to as ‘computer’ keyboards. 

Hard copy, non-counting keyboards 

These keyboards are generally counting keyboards with the counting 
mechanisms removed, or else typewriter systems that were originally 
intended for other reasons and are now being offered to the Graphic Arts 
Industry. They cost more than blind devices, and do not permit the same 
high keying speeds due to the necessity of periodically returning the carriage 
which breaks the operator ryhthm. Their primary application is the more 
accurate keying of straight matter, or the training of new keyboard 
operators. 

Once the material has been typeset and reviewed by the proof reader and the 
appropriate areas of the action identified, the information is passed on to the 
place where the material is corrected. 


Irrespective of the nature of the error, it will always be corrected by one of 
two actions: (a) something must be inserted, and-or (b) something must be 
removed. With these two correction steps in mind, we now need to address 
the reason for the change. 

This involves the correction of keystroke errors, or spelling mistakes. To 
implement the correction, it is necessary to have access only to that portion 
of the material in question. There is no need to be able to review either 
preceding or subsequent material. Further, the material can be corrected by 
the substitution of very small amounts of new data; usually the insertion of a 
correct letter or correct word is all that is required. 

This involves the complete overview of all the material, or at least the 
overview of some complete segment of it. Editing may require entire 
paragraphs removed, replaced, or relocated elesewhere in the sequence. To 
implement revisions, it is necessary to have access to large amounts of 
material, and the amounts of new data may be considerable. 

This applies to such problems as updating telephone directories, personnel 
rosters, guidebooks, etc. The material in question is typically handled in a 
unit (e.g. name and address, name and department number-telephone, etc.) 
and whole units of information are removed or inserted. Access to large 
amounts of the material is not required, since the information is usually 
arranged in sequence (alphabetic or numeric) and the insertions or deletions 
can be structured the same way. 

The factors that influence corrections are: amount of material to which 
access is necessary, size of insertions-deletions (number of characters or 


Composition Input 47 



words), method used to insert material, time permitted to complete 
alternation. 

The process of correcting, editing, and-or updating material is inherent in 
any typesetting operation. It is useful to understand the traditional 
procedures before examing some of the newer devices being offered. Since 
proof reading is virtually the same in all cases, our primary focus will be on 
the manner in which corrections and-or alterations are handled. 


Manual Methods 

Proofreading is done from the galley. The galley may be the output paper 
from a phototypesetter, or it may be a reprocopy pulled from the chase 
containing the cast lead slugs. After proof reading, the fix can be 
accomplished in two ways: 

Hot Metal 

The correct information is keyed, and new slugs or lines are cast. These slugs 
are inserted in the chase in place of the incorrect or non-valid lines. The 
minimum amount of material that can be replaced is one full line. 

Phototypesetters 

The correct information is keyed, and the tape run through the 
phototypesetting machine. The output paper is then pasted over the incorrect 
or non-valid portions of the original galley. In practice, all the corrections 
are keyed and processed at the same time, so the output paper may contain a 
number of corrections. 

If the correction involves the insertion of some information previously 
omitted, it may be necessary to rekey entire paragraphs of correct 
information because of the ‘domino effect’ of the insertion. 

With either method, it is impossible to proof read until after the typesetting 
function. In the event of a number of errors, the usual procedure is to re-key 
all the information and discard the first pass. The lack of visibility of the 
data prior to typesetting often results in a significant amount of waste, in 
terms of people time, machine time, and supplies cost. 


Automated Methods 

A number of systems are available today, offered in different configurations 
and with varying degrees of flexibility. Basically these machines offer the 
same principal thing; the facility to display and massage the material prior 
to typesetting. The methods used to display and the flexibility of the 
alteration procedure are the primary considerations when evaluating the use 
of these systems. Some machines permit proof reading and correction to be 
done at the same time, other systems are structured so that correction takes 
place on a second pass. Some systems can address several different types of 
alterations. 


48 Composition Input 



It is very difficult to make meaningful comparisons of different types of 
systems, due to the varying capabilities of the hardware and the extremely 
broad range of correction - editing - updating problems. Rather than make 
comparisons, it seems more reasonable to illustrate the principle of a given 
system as applied to a given alteration task. The best system for a given 
application depends on the application and the economics governing it. 

Keyboard methodology 

The earliest tape perforating keyboards were completely mechanical. On a 
mechanical keyboard the keys are linked to the tape punch by a series of 
levers. Under each key, and running from the front to the rear of the 
machine are coding bars which are serrated in accordance with the code to 
be punched, and these operate, a series of combination bars which run at 
right angles to them. There is a combination bar for each code channel, and 
they interpose a system underneath the punch knives. When a key is 
depressed, the required code is set on he interposes under the punch knives 
and at the same time a clutch is operated which causes a continuously 
running electric motor to operate a cam under the punch knives. This cam 
forces the knives through the paper tape in those channels where the 
interposes have been set on the code bar. 

There is necessarily a ball-lock device on the mechanical keyboard to 
prevent two keys from being depressed, simultaneously. This can also be 
adjusted to operate early or late on the keystroke. With the continuously 
running motor, the mechanical movements make the keyboard noisy in 
operation. Because of the type of construction involved, the slope of the key 
panel of this type of keyboard is usually much greater than the 12 or 15 
degrees which has been found to be optimum on modern high-speed typing 
devices. Also, the space between keys is generally 7/8”, as against the 3/4” 
spacing of typewriters. Because of the mechanical design it is impossible to 
change the layout of the keyboard or to add additional keys, so that these 
machines are very inflexible. 

In an effort to overcome the problems of slow and heavy operation of the 
completely mechanical keyboard, electromechanical keyboards were 
introduced. These keyboards still operate coding bars from the keys, but the 
combination bars of the mechanical keyboard are eliminated. The operation 
of a key allows the code bar to slide forward and operate on a series of 
contacts. These contacts are wired to the code magnets of the punch which 
set up interposers under the punch knives. The key also operates a clutch 
which allows the continuously running electric motor to turn a cam and force 
the punch knives through the paper. The mechanical interlock has been 
taken off the actual key and operates on a sliding code bar on some types of 
these machines so that only one bar is allowed to make contact at one time. 


The third type of machine is the contact type. It employs a simple electrical 
contact or switch which, when the key lever is depressed, brings two pieces of 
electrically conductive material into contact, allowing a current to pass. For 
the contact to break when a key is released, it is usual for a contact to be 
mounted on very spring-responsive material. Because of this, whenever a 
contact is made by depressing the key, ‘contact-balance’ occurs. That is, the 


Composition Input 49 



two contacts make and break contact rapidly. This causes two problems: 
first, balancing-contacts can allow a code to pass after the key has been 
released, and this can lead to a corrupted code being punched when the 
following key is operated quickly. Second, whenever a current is broken by 
means of a contact or switch, arcing takes place. To a lesser extent, arcing 
also occurs when the contact is made. In time, the arcing builds up a high- 
resistance deposit on the contacts and thus reduces the current flying 
through the contacts. 

To overcome the known problems of switching current, a photo-electric 
keyboard was produced by the Invac Corporation. This type of keyboard 
consists of a number of light channels with a photo diode at one end and a 
light source at the other. There is a light channel for each bit of the code plus 
a control channel, and they run parallel to the back of the machine. Over the 
light channels and at right angles to them are a series of code bars which are 
operated when a key is depressed. These code bars are designed in such a 
way that a shutter drops into the light channel when the key is operated, thus 
preventing the light from reaching the photo diode at the other end. The 
number and position of the shutters on the code bar determine the code to be 
punched by that key, since the photo diode only passes current when it is 
illuminated. The count is used to actuate the punch knife in that particular 
bottom of its stroke, the shutter would not reach the bottom of the light 
channel. This would allow light to pass through and cause a wrongly 
punched code. To meet this intrinsic problem in the photo-electric 
keyboards, the keys are power-assisted so that when a key is depressed a 
solenoid or electromagnet is operated, it pulls down a key plus its code bar, 
thus insuring that the shutters reach the bottom of the light channel. 
Unfortunately the force with which the shutters hit the bottom of the light 
channels causes them to vibrate, thus allowing spurious light onto the photo 
diode. This spurious light can, of course, cause mispunching when the 
keyboard is operated quickly. 

In an effort to overcome this, a very heavy mechanical interlock had to be 
introduced, and its slows the keyboard down considerably. Consequently, it 
can be seen that ‘contact-balance’ has been replaced by ‘shutter-flutter’ with 
the same damaging results. Another cause of trouble on the photo-electric 
keyboards is the gradual reduction in the strength of the light source on the 
output from the photo diodes. This can eventually lead to mispunching. As 
with mechanical and electro-mechanical machines, there is no flexibility of 
keyboard layout, and if additional keys are required, they have to be built 
into a supplementary keyboard. 

Finally, there is the reed switch type of keyboard. It incorporates keys 
consisting of a relay operated by a magnet. When the key is depressed, the 
magnet is moved down the reed and causes the contacts of the reed relay to 
close and so allow current to pass. When the original keyboard of this type 
was introduced some years ago in Europe it was thought that here at last 
might be the answer to the problems on mechanical, electro-mcchanical, 
contact and photo-electric keyboards. However, the initial installation was a 
failure because the operators complained emphatically about the feel of the 
keyboard. An operator likes to know that when he or she presses a key that it 
has operated effectively, and this can only be verified by a feel that 
something has happened at some some point in depressing the key. On a reed 


50 Composition Input 



Dual Image Keyboard 


Dual Image keyboard devices must be considered in a separate category, due 
to the difficulty in comparing them to any traditional keyboards. Dual 
Image offers the speed advantages of blind noncounting keyboards, 
combined with all the visibility benefits of hard copy machines. 

The system utilizes electronic keyboard elements which are solid state 
devices. With the exception of the key shaft itself, they have no moving 
parts. A two character memory is provided so that burst keying does not 
result in any lost or erroneous characters, and/ this is coupled with a 
recording mechanism that can accept data at speeds up to 30 characters per 
second. The keyboard, the memory, and the recorder combine to form a 
keyboard system that can operate at keying rates in excess of 100,000 
keystrokes per hour. 



The main keyboard of the Dual Image Recorder. 


The use of printed paper tape gives all the benefits of hard copy, yet does not 
introduce any of the disadvantages of typewriters. All other hard copy 
keyboard systems depend upon the use of a typewriter for visibility,which 
has two major disadvantages: 1) limited character set, 2) need for carriage 
return, which breaks keying rhythm and reduces effective keying speeds. 


FInally,.pass.It.on.to.your.machines.at.high.speed. 

** * ,, •• ji •j 1 i.. i «•*. B • ■•* • , . i.f 


Dual Image tape. 


Composition Input 51 









- i 


'llllSlll: 


Characters and functions are immediately visible 
when keyed on the Dual Image keyboard. Dual Image 
is an interesting approach to the hard copy concept. 


Dual Image is not affected by either drawbacks, since the machine can 
display a full set of 128 different symbols. The printing is on a continuous 
strip of tape, so the operator never needs to break keying rhythm. 


It can display both upper and lower case textual material, just like a 
typewriter. In addition, the expanded character set permits control codes to 
also be printed, and these control codes describe the function so that the 
operator needs to make no 'meaning 1 translation. For example, Quad Left is 
QL, Upper Rail is UR, Lower Magazine is LM. Other hard copy machines, 
using typewriters, frequently print nothing for the command codes. 
Sometimes a command code is a textual character printed in red. 




52 Composition Input 








The heart of the Dual Image system is the high speed impact printer. The 
tape to be printed is positioned between a hammer and a high speed print 
wheel (which rotates at 1800 rpm). When the desired character is in position, 
the hammer strikes, printing both the human readable symbol and the 
machine readable code at once. There can never be any disagreement 
between what the human reads and what the machine reads. The last 
character printed is immediately visible to the operator, so that material 
being keyed can be immediately reviewed and confirmed. The recorder 
operates with a minimum of moving parts, ensuring a long life and trouble 
free operation. 


Once the tape is prepared, it can be easily read into a computer or 
phototypesetter by using a Dual Image reader. This is a solid state photo- 
optical scanner, and can read information at rates up to 300 characters per 
second. As the tape is moved past the lens, the bit pattern is focused on a 
plane of photo-transistors, and the character is electronically read. The 
reader can be made to emulate a variety of traditional devices, such as the 
Teletype CX 100 reader, the Tally 424, or Digitronics readers. 


CONTROL 

AREA 



ER 


DR 


CR 


BR 


AR 


Fig. 1 Proposed U.S.A. Standard: Logical Bit Pairing 


CONTROL 

AREA 



the typewriter key set as used for computer input. The 
program was sponsored by the American National 
Standards Institute (ANSI) and the Business Equip¬ 
ment Manufacturers Association (BEMA). 


Composition Input 53 











































































































How to tailor a keyboard. Here is one company that 
makes the job fairly simple. Invac provides a layout 
grid for indication of keytops and codes. 


Combination Chart Used for All 
INVAC Keyboard Formats 

INVAC Corporation provides three basic 
keyboard formats: 

(1) PK-244 (48 keys) 

(2) PK-264 (64 keys) 

(3) PK-275 (75 keys) 

Three shades of gray indicate the limits 
of each keyboard format (see Legend 
below). Note that the PK-275 extends 
across all three shaded areas , the PK-264 
uses the two lighter shades, while only the 
central, lighter area comprises the PK-244. 

Each key position is connected by a line 
to a row of code and key color specification 
spaces in the lower section of the chart. 
Please note that individual "key-numbers" 
identify both the key position and the 
related code and key color data. 

Please refer to the notes on this page 
and the "Check List" on Page 1. 

A sample chart on Page 4 shows a typi¬ 
cal customer specification. 

Other Information Needed to Complete 
the Specification 

A step-by-step check list has been provided 
on page 1 to facilitate specifying the key¬ 
board format and code data. Space has 
been provided on page 4 for additional 
remarks. 

Notes: 

1. Three keyboard formats are available: 48 
keys, 64 keys, and 75 keys. In addition, 
certain keys may be arranged in either 
diagonal rows (as shown) or vertical rows 
(such as one finds on an adding machine). 
Unless otherwise specified the diagonal 
keyboard will be provided. See Table 1, 
page 1 for special key size. Please enter 
special key size in Check List under 
“Special Features”. 

2. Use key numbers for any special refer¬ 
ences such as key sizes, colors, special 
contacts, etc. 

3. For convenience, and to minimize error, 
please designate code bits by repeating 
bit number in the appropriate space. See 
illustrated example on page 4. 

4. See Table 2, page 1 for available keytop 
colors. Unless otherwise specified. Dark 
Blue (“B") keytops will be provided. Please 
enter this data in the appropriate spaces 
provided on the Format Chart. 

5. See Table 3, page 1 for available keytop 
fill colors. Unless otherwise specified, 
White ("W") keytop fill will be provided. 
Please enter this data in the appropriate 
spaces provided on the Format Chart. 


FORMAT CHART for INVAC Keyboards, Codes, and Keys. 



LEGEND FOR KEYBOARD FORMATS 


mu + 

n + 

= PK-275 Keyboard 

n 

jj + 

= PK-264 Keyboard 



= PK-244 Keyboard 


54 Composition Input 


Code and Key Data (See notes designated by superscript numbers) 





























4. Input media and coding 


The Greeks were first again. About 300B.C., Polybius reported the following 
system: stations were erected at many locations which consisted of two walls 
about seven feet long and six feet high, separated by a space of three feet. At 
night, one or more torches, as needed, but no more than five, were placed on 
top of the walls. Certain combinations of torches represented Greek letters. 
Two torches on the right wall and three on the left may have stood for the 
letter 4 H’ as an example. Thus, words were spelled out letter by letter on this 
five-unit, center-feed “tele”-torch communication system. 

In 1887 Herman Hollerith constructed the first electromechanical system for 
recording, computing and tabulating digital data, which he then used to 
record the 1890 census. Holes were punched in cards with a conductor’s 
punch. These cards were positioned over a series of mercury-filled cups and 
at the touch of a lever, telescoping pins projected to the card’s surface and 
then through, if there was a hole. The pin reaching the mercury completed an 
electrical circuit and and this in turn moved a pointer one position on a dial. 
The punched (or punch) card was one of the first methods of recording 
information. 

A punched card measures 7 3/8 by 3 1/4 inches and is 0.007 inches thick. It 
contains 80 columns (the 90 column card developed by Univac was 
discontinued in 1966) which are numbered from left to right, 1 to 80. 
Vertically there are twelve rows numbered 12, 11, and then 0 to 9, from top 
to bottom. The two top rows are also called the X(11) and Y(12). The upper 
left-hand corner may be cut, or the corners rounded, depending on the 
system utilized. An alpha or numeric character is represented by holes 
punched in one or more locations of a single row. Groups of characters, or 
rows, such as columns 1 through 30 or 23 through 34 for example, are called 
fields. A field is a unique group of characters and may represent a part 
number or a description or an address. Thus all addresses may be put in 
certain columns, and we always know (as does the computer) where to find 
the address for comparison, sorting or correction purposes. 

The punched card idea goes back to the early 1800’s. Punched cardboard 
patterns were used to direct textile looms by mechanically selecting hooks 
which raised the longitudinal threads to make a passage for the shuttle, 
which “set” the crossthreads. J.M. Jacquard thus automated the weaving 
process and paved the way for the automatic tape operation of industrial 


Composition Input 5 5 



machinery that would come 150 years later. The punched hole was also used 
by the Monotype as a tiny “valve” which controlled air pressure and in turn 
positioned a matrix case. 

Punched cards are an almost permanent recording medium. They may be 
retained for long periods and thus eliminate re-punching. Each card is a 
“slice” of information, a record, and may be changed without affecting other 
cards. The punched card began the unit record concept. 

The early telegraph, from 1850 to 1920, used the Morse code of dots and 
dashes. The duration of a dot was one unit; that of a dash was three units. 
There were also three units between characters and six units between words. 
J.M.E. Baudot applied the idea of two early telegraphic developers, Gauss 
and Weber, of a code using five units to represent the alphabet, numerals and 
special symbols. This system, with some variations in character allocations, 
became the standard telegraphic code system. A seven unit code was devised 
by J.B. Moore which permitted detection of code mutilation by transmission 
and reception devices. A modified form of the seven unit code was later 
expanded by ELC.A. VanDuuren. 

In 1858 Sir Charles Wheatstone employed a perforated tape to operate the 
telegraph transmitting mechanism. At first it was driven by a clock device 
and later by an electric motor. The Morse signals were received by an 
“inker” in which an inked wheel marked the dots and dashes on a moving 
tape. 



A 

a 

B 

b 

C 

c 

D 

ri 

E 

e 

0 

G 

e 

0 

B 

e 

1 

B 




B 

b 

i 

S 

s 

B 


B 

B 

B 

B 

i 

z 

fn 

1 

w 

2 

1 

B 

1 

1 


1 

ft 

9 

B 

B 

B 

B 

© 

B 

B 

B 

nr 

' 3 

Em Ld| 

3 

c 

UJ 

CL. 

in 

C 

UJ 

Cl. 

m 

B 

UJ 

o- 

(Si 

j=. 

3 

cb 

cL 

| 

i 

>■ 

Jt 

UJ 


X 

’S) 

g 

I 

CtL 

| 

E 

















3 


3 






3 

3 


□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 


• 


■ 

□ 

m 

"5 

□ 

□ 

□ 

■ 

i 

□ 

□ 

■ 

□ 

□ 

□ 



■ 

□ 

□ 

■ 

81 


■ 

■ 

□ 

■ 

□ 


• 


□ 

□ 

□ 

□ 

□ 

□ 

□ 

■ 

■ 

□ 

□ 







w 

_ 

□ 

□ 

■ 


V 

3 

□ 




• 





2 

□ 


□ 

■ 



• 


□ 

□ 

□ 

□ 



■ 

□ 

□ 

□ 

■ 


□ 

□ 

□ 


■ 

■ 

■ 

□ 

■ 

□ 


■ 

□ 

□ 


□ 





□ 

□ 


□ 


□ 

□ 



■ 

m 

□ 

& 

□ 

□ 

□ 

□ 

BB 

□ 

□ 



Q 

o 

0 


0 

0 

O 

o 


o 

3 


L° 

0 

Ti 

° 


3 

0 

0 

3 

0 

o 

o 


0 





0 

3 

° 

• 

o 

° 

o 

o 

o 

o 

3 

o 

0 


o 


0 

o 

0 

° 


° 

° 

0 

0 



3 

- 

0 

o 

3 



□ 

■ 

■ 

□ 

■ 

□ 

□ 


□ 

■ 

□ 

□ 


□ 

□ 

■ 

□ 


□ 

□ 


□ 

□ 


□ 

■ 

■ 

r - 

■ 

□ 

□ 

□ 


□ 

■ 

□ 

□ 

■ 

m 


■ 

■ 

□ 

□ 

□ 

□ 

■ 

□ 

□ 

□ 

m 

■ 


BB 

□ 

■ 

■ 



4 


□ 

□ 

□ 

■ 

□ 

□ 

■ 

■ 

□ 

□ 

_ 

□ 

□ 

□ 

■ 

■ 

□ 

■ 


■ 

□ 

■ 

□ 



□ 



□ 

■ 

■ 



□ 

■ 

□ 

□ 

□ 

□ 

□ 

■ 


■ 

■ 

■ 

□ 



■ 

■ 

□ 

□ 

BB 

■ 

□ 

□ 

□ 

□ 

□ 

■ 

5 


□ 

■ 

■ 

■ 

■ 

□ 

□ 

_ 

□ 

_ 

□ 

□ 

_ 

□ 

□ 

□ 

_ 

_ 

□ 

_ 

□ 

□ 

□ 

□ 

□ 

□ 

□ 


■ 

□ 

□ 



□ 

□ 

U 


U 

U 


U 

□ 

□ 

□ 

□ 

U 


* 

■ 

u 

□ 

■ 



□ 

□ 

□ 

□ 

3 

□ 


The TeleTypeSetter code. 


In 1915 Western Union established the multiplex system of printing 
telegraphy. It provided a number of independent channels of communication 
going in both directions at the same time on a single wire. Here, punched 
tape was used extensively for the first time. The Baudot five unit alphabet 
was used primarily. Five impulses of negative and positive current were 
combined to form a given letter. Thus A was represented as two positive and 
three negative units. 

Punched tape may be paper or plastic. In either case there is always one long 
lengthwise row of small sprocket feed holes and five, six, seven or eight rows 
of larger parallel holes which represent character and function codes. Here 
are some of the types and forms of punched paper tape: 

Oiled: tape impregnated lightly and uniformly with oil for lubrication and 
ease of punching. 

Dry. paper tape, period. 

Mylar, plastic tape which is more durable and may be re-run more often 
Strip: two to four foot sections of tape. 


56 Composition Input 






















Fan-folded: tape folded every six or so inches into a stack. 

Roll: about 700 feet of tape in its most used format. 

Center-feed: the sprocket holes are lined up with the middle of the code 
holes. 

Advance-feed', the sprocket holes line up with the leading edge of the code 
holes. This approach was developed to avoid confusion with which end goes 
first. When the sprocket hole is to the left of the code hole (on a right to left 
reader) the tape is being read correctly. 

Magnetic tape is polyester plastic with a coating of magnetic particles. 
Invented by O. Smith in 1880, who impregnated a cotton thread with steel 
dust, mag tape took fifty years to perfect. Today’s tape is about one half inch 
wide, one half mil to one and a half mils thick and 2400 feet long. Quarter, 
three quarter and even inch tape in various lengths is also used. 

Information is recorded on the tape by magnetizing narrow strips called 
tracks in alternating directions. Thus, one frame of paper tape with holes 
representing binary codes is represented on mag tape by the presence of a 
magnetic flux reversal; no holes are represented by the absence of a flux 
reversal. 200,556, or 800 bits per inch (bpi) or code frames may be recorded. 
The width and bpi of mag tape is determined by the tape reader employed. 
Tracks are presently designated as either seven-track or nine-track, and each 
apply only to certain computers. A mag tape cassette is a length of magnetic 
tape wound so as to form a continuous loop, with an opening at which the 
tape may be read and recorded. 


Each hole in a card or tape or flux reversal on mag represents a bit. Bit 
stands for binary digit and means yes or l or on if it is there, and no or zero 
or off if it is not. Here’s where a little arithmetic comes in. We use a 
numbering system every day that is based on the number 10. In the number 
567 the “7” position is digits (total: 7); the "6” position is tens (6 x 10 equals 
60); the “5 position is hundreds (5 x 100 equals 500); so the number is 
expressed as 567. Thus we count by multiplying the number of tens in the 
position in which a number occurs. Going from right to left: 


10 6 

10 5 

10 4 

10 3 

10 2 

1,000,000 

100,000 

10,000 

1,000 

100 


V-I-P 

CODE 

V-I-P 

FUNCTION 

5 4 3 2 1 0 

• • • • 

EM LEADER 

• •*••• 

EN LEADER 

• • 

SPACEBAND 

• • • • 

EM SPACE 

• • • • • 

EN SPACE 

• 

THIN SPACE 

•••••• 

UNSHIFT 

• • • •• 

SHIFT 

• • • • 

SUPERSHIFT 

• • • •• 

QUAD LEFT 


QUAD RIGHT 

••••• • 

QUAD CENTER 

• • • ••• 

LOWER RAIL 

• • • •• 

UPPER RAIL 

• • ••• 

BELL 

• • 

ELEVATE 

• • • 

ADD THIN 

• • 

RETURN 

• • • 

PAPER FEED 

••••••• 

RUBOUT 

• 

• 

TAPE FEED 


Command codes for the Mergenthaler V-I-P. 


10 1 
10 


(10 0 ) 
( 1 ) 


Composition Input 5 7 



The above review is more than you need to know about binary arithmetic (so 
why tell me after I read it?); but it is a necessary introduction to coding. To 
encode only the numerals 0 through 9 would require codes four bits long: 
0000= 0, 0001 = 1,0010= 2, 0011 = 3, 0100= 4, 0101= 5, 0110= 6, 0111 = 
7, 1000= 8, 1001 = 9. Actually sixteen combinations are possible; not nearly 
enough to encode an entire alphabet. A five bit code, such as the teletype 
code, permits 64 possible combinations. Still inadequate for typesetting. The 
teletypesetter utilizes a six bit code with 128 possible combinations. 
Advanced typesetters require even more codes than this. 



TAPE 

CHANNEL 

NUMBERS 

CHARACTER 

OR 

FUNCTION 

Dl 

a 

a 

ai 

Q| 

Q 

tu 

Ell 

SI 

KM 

UNSHIFT 

SHIFT 


■ 

a 

i 

s 

a; 

fli 

Eli 

Em 





mu 

wrm 

mom 

• 

ZZ 


zz 

a 

A 



■ 

mom 


mm\ 

hhi 

Oil 

Ofll 

b 

B 




g 

Efl 

Bi 

Dll 

EMI 


c 

C 



g 

mam 




EMI 


d 

D 



VMI 

mrm 


* 




e 

E 



■HI 



mfli 

Efl 

OH 


f 

F 




g 

km 

♦ 


Dll 

EMI 

9 

G 






KM 

log 


• 

h 

H 




gg ! 

ii 


Dl 



i 

1 



ggpj 

«bH 

worn 

^Z 

g 

EH 


i 

J 



mm 

worn 

MTi 

mg 

g 

OR 


k 

K 





i 

^Z 

J 

Hi 

OR 

1 

L 






H 

Efl 

mm 

EH 

m 

M 




g 

RH 

gOM 


OR 

H 

n 

N 




g 

■ 

^z 


wrm 

g g 

o 

O 



~ 

g 

Efl 

mm 

IDfl 

IB 

KH 

p 

P 



■ 

g 

Efl 


• 


• 

q 

Q 



— 

■ 


* 


• 


r 

R 



mm 

mrm 


* 

• 



s 

S 




i&ig 


» 


mm 

KH 

t 

T 




mrm 

worn 

* 

• 

zz 


u 

U 





mm 

m 

Efl 

EH 

KH 

V 

V 




• 

mm 

g 



KH 

w 

W 



■H 

m 


♦ 

wrwwt 

KM 

ROH 

X 

X 






♦ 

mm 

ggg 

wrm 

y 

Y 



in 

mrm 


• 



KH 

2 

Z 



man 

mrm 


9 

mm 

EM 

!■ 

i 

'/. 



u 

• 

mm 

• 


HHI 

KH 

2 

% 



mrm 



* 


zz 


3 

% 



mrm 


mm 

♦ 

WWM 

wrm 


4 

'A 



mom 



♦ 



arm 

5 

y. 



mrm 

mom 


* 

mm 


KH 

6 

3 A 



mom 


mm 

«r 

■fli 



7 

IHflHI 



mom 


mom 

♦ 

EM 



8 

IHSEEHl 



arm 



* 


KM 

arm 

9 

& 



mom 

- -'— 

mm 

9" 

wrm 


mrm 

0 

? 



mrm 



9 

mom 

worn 

KH 

. 

BBS 



mrm 



mm 

worn 

mm 

flflfl 

COMMA 

flggQQZflHI 



mrm 


• 



wrm 

mrm 

; 

: 



■EX 

mom 

mm 

• 


zz 


$ 

! 



mom 

mm. 

mm 

9 


mmm 

mum 

) 

( 



worn 



mm 


KM 


APOS. 

QUOTE 



mom 

■OH 

■HR 



'KM 


HYPHEN 

+ 




mtm 

ggg 

r ~ 

mmm 

ggg 


TAPE 

FEED 



SS 

— 

arm 

gH 

mom 

ROM 


EN SPACE 



mom 

flOP 

mrm 

igH 

arm 


arm 

EN LEADER 



mom 

flOP 

Hi 

♦ 

mom 

iggg 

flH 

EM SPACE 




■mi 


• 

mom 

iggg 

KH 

EM LEADER 



BOR 

:ggg| 

\mm 

mm 

mmm 

Iggg 

IKM 

VERT. RULE 



mom 

mm 

mm 

mm 

1 |ggg 

IMM 

IRHRg 

THIN 

SPACE | 






ggg 

imm 

Iggg 

rz 

SPACE BAR 






ggg 

;PSgi|* 

IKM 

IRHHH 

RETURN 





mom 


I ggg 

rz 

IRHfl 

ELEVATE 



mrm 

mm 

wrm 

SS 

imm 

iggg 

IflH 

PAPER FEED 



mm 

in 

Efl 

Bi 

■ihhi 

IKM 

I ROH 

SHIFT 




mm 

Efl 

ggg 

IKfl 

IKM 

in! 

UNSHIFT 



SS 

arm 

HI 

mm 

IHMC 

IKfl 

IROHI 

UPPER RAIL 



mom 

mrm 

Efl 

Hi 

IMM 

IKfl 

IflH 

1 LOWER RAIL 



mom 

mom 

Kfl 

flH 

IH1M 

IKM 

IRHfl 

STOP 



r *i 

mom 


ggg 

I MTM 

IKfl 

iwnm 

RUB 

OUT | 





ggg 

ggg 

Iflfll 

Iflfll 

IRHfl 

ADD THIN 




SS 

mmm 

ggg 

IKfl 

in 

Imm 

QUAD LEFT 



mom 

flH 

Kfl 

H 

IKfl 

IKS 

in 

} QUAD 

CENTER j 



ggjI 

mom 


Hi 

IKfl 

IKfl 

IHH 

QUAD RIGHT 


• 



8 

ggg 

yaggggg 

f mmm 

Iflfll 

IRRRg 

* ADD 7 

• 


s 

Hi 


MM 


Iflfll 

iflHH 

ADD 8 


Coding chart for the Friden 8201 keyboard. 


58 Composition Input 







































Because the majority of phototypesetters standardized on the TTS 6-level 
code due to the proliferation of keyboards, newer devices require multiple 
codes to access certain characters and functions. Thus, some keyboards 
produce six-level tape with a seventh or eighth bit key to make no code more 
than two keystrokes. Others produce 8-level tape which provides a more 
complete code array. 


Coding chart for the Friden LCC-VF keyboard. 


TAPE 

CHANNEL 

NUMBERS 

CHARACTER OR FUNCTION 

6 

5 

4 

FEED 

3 

2 

1 

UNSHIFT 

SH 

FT 

Lower Rail 

Upper Rail 

Lower Rail 

Upper Rail 


• 

• 

• 




a 

a 

A 

A 


• 




• 

• 

b 

b 

B 

B 



• 


• 

• 


c 

c 

C 

C 


• 




• 


d 

d 

D 

D 


• 


• 




e 

e 

E 

E 


• 


• 

• 

• 


f 

f 

F 

F 



• 

• 


• 

• 

9 

9 

G 

G 




• 

• 


• 

h 

h 

H 

H 



• 

• 

• 



i 

i 

1 

1 


• 

• 

• 


• 


i 

i 

J 

J 


• 

• 

• 

• 

• 


k 

k 

K 

K 



• 

• 



• 

l 

1 

L 

L 




• 

• 

• 

• 

m 

m 

M 

M 




• 

• 

• 


n 

n 

N 

N 




• 


• 

• 

o 

o 

O 

O 



• 

• 

• 


• 

P 

P 

P 

P 


• 

• 

• 

• 


• 

q 

9 

Q 

Q 



• 

• 


• 


r 

r 

R 

R 


• 


• 

• 



s 

s 

S 

S 




• 



• 

t 

t 

T 

T 


• 

• 

• 

• 



u 

u 

U 

U 



• 

• 

• 

• 

• 

V 

V 

V 

V 


• 

• 

• 



• 

w 

w 

W 

w 


• 


• 

• 

• 

• 

X 

X 

X 

X 


• 


• 

• 


• 

y 

y 

Y 

Y 


• 


• 



• 

z 

z 

Z 

u 


• 


• 

• 

• 

• 

1 

V 

ffi 

w 

• 

• 

• 

• 



• 

2 

B 

ffl 

M 

• 

• 


• 




3 

C 

3 


• 


• 

• 


• 


4 

D 

* 

0 

• 



• 



• 

5 

E 

ff 


• 

• 


• 

• 


• 

6 

F 

se 

K 

• 

• 

• 


• 



7 

G 

& 

N 

• 


• 


• 



8 

R 



• 



• 


• 

• 

9 

T 

fi 


• 


• 

• 

• 


• 

0 

J 

fl 

Y 

• 



• 

• 

• 

• 

Period 

Period 

Period 

Period 

• 



• 

• 

• 


Comma 

Comma 

Comma 

Comma 

• 


• 

• 


• 

• 

Semicolon 


Colon 

S 

• 

• 

• 

• 




$ 

P 

j 


• 

• 


• 



• 

) 

0 

( 

A 

• 



• 


• 


Apos./Quote 

Apos./Quote 

Quote 

Quote 

• 

• 


• 


• 


— 


06 

H 




• 




TAPE 

FEED 



• 


• 

• 

• 

• 


EN SP. 

EN LD. 



• 

• 

• 

• 

• 


• 



EN LD. 

L 

• 

• 


• 

• 



EM SP. 

EM LD. 



• 



• 

• 


• 



EM LD. 

EM SP. 

• 


• 

• 



• 

VERT. RULE 

l 

VERT. RULE 

1 


• 

• 

• 


• 

• 



SHIFT 

SHIFT 


• 

• 

• 

• 

• 

• 

UNSHIFT 

UNSHIFT 



• 

• 


• 


• 

• 

UPPER RAIL 

• 

• 

• 

• 


• 

• 

LOWER RAIL 

• 

• 

• 

• 


• 


STOP CODE 

• 

• 

• 

• 

• 

• 

• 

CODE DELETE 

• 

• 


• 

• 

• 


r QUAD LEFT 

• 


• 

• 

• 

• 

• 

K QUAD CENTER 

• 

• 

• 

• 

• 

• 


QUAD RIGHT 

• 



• 




THIN SPACE 




• 

• 



SPACE BAR 




• 


• 


CARRIAGE RETURN 



• 

• 




ELEV. ELEV. 

• 


• 

• 




PAPER FEED 

• 



■- 

• 

• 



SPACE BAR and THIN SPACE 

0 

1 

2 


3 

4 

5 

HOT METAL EQUIVALENT 


Composition Input 59 









VALUE UNSHIFT 5 4 3 2 10 SHIFT SUPER SHIFT 


0 0 





"IF 

r 

■n 


z 



0 1 

nr 







■ 

□ 

HiH 


III 

c 






■ 

u 

p 

E 


IH 

3 






i 

□ 

a 

3/8 

3 

ID 







• 


■ 

e? msm 


EQ 

ADD POINT 






• 


□ 



m 

a 






• 

\9 


A 


ran 

f 






• 

9 


mmm h 

I 

m 






n 




»&m:-nnii 


HI 


■ 




□ 


■ 

□ 

[KILL LINE 


1 0 

s 





LJ 


E 

■ 

S 


1 I 

□ 





n 


□ 

LJ 

□ 


1 2 

i 

■ 

■ 



u 

□ 

■ 


I 


13 

8 

■ 

■ 



u 


■ 

□ 

EM- 

8 

1 4 

u 





u 

M 

LJ 

■ 



1 5 

7 





□ 

LJ 

□ 

□ 

Sgsii 

7 

1 6 

RETURN 




□ 

■ 


■ 

■ 



1 7 

APO S » 




n 



__ 

□ 

* 


1 8 

d 




u 



LJ 




1 9 

HY- 




u 



LJ 

U 



2 0 

r 

■ 

■ 


u 


□ 



R 


2 1 

4 


z 


u 


u 


u 

'/2 

4 

2 2 

3 

t§ 

■ 


u 


u 

• 


J 


2 3 

COMMAND 




u 


u 

U 

u 

COMMAND 


2 4 

n 




□ 

□ 



z 

N 


WTW 

COMMA 




n 

□ 



□ 

1 1 1 11 


VY1 

1 




lj 

[J 


□ 


F 


HI 

Q L 




a 

□ 


□ 

W 

a l 


2 8 

c 

■ 

■ 


u 

□ 

□ 

■ 


c 


HI 





u 

u 

u 


u 



fn 

k 




u 

u 

u 

□ 

z 

K 


3 1 

Q R 




□ 

□ 

□ 

□ 

□ 

Q ft 


32 

t 



• 






mamm 


KOI 

5 



r 9 





□ 

mum 

S 

III 

z 



• 




□ 

■ 

z 


ni 

r 



□ 




□ 

□ 

1 

*/. 

HI 

i 



U 



□ 


z 

L 


EDI 

SUPER SHIFT 



• 



u 

■ 

□ 

SUPERSHIFT 


III 

w 



□ 



u 

u 


W 


fDI 

2 



u 



□ 

□ 

□ 

'/4 

1 

eh 

h 



u 


LJ 




H 


4. 

EM LDR 



□ 


□ 



• 

EM LOR 


WWW 

Y 



n 


n 


n 


Y 


EH 

6 



□ 


• 


□ 

□ 

i U 

i 

wrm 

P 



□ 


• 

• 



P 


EH 

0 ZERO 



u 


• 

• 


u 

? 

o tua 

EH 

q 



• 


• 

• 

• 


Q 


EH 

EN LDR 



•1 


• 

• 

□ 

□ 

EN LOR 


EH 

O 



• 

• 




z 

O 


EH 

9 

■ 

m 

□ 

• 




□ 

& 

• 

EH 

b 



• 

• 



□ 

z 

B 



U R 



• 




u 

□ 

UR 



9 



• 

• 

• 


• 

• 


z 

6 



• 

» 





□ 

; 

4 


SHIFT 

11 

m 

□ 



• 

□ 

z 



EH 

L R 



• 

•1 


• 

u 

LJ 

LR 


EH 

m 



• 

•i 





M 


EH 

PERIOD 

■ 

■ 

□ 





• 

PERIOD 


EH 

X 

z 


• l 

• 

M 


• 


X 


EH 

1 

■ 

■ 

□ 

• 

M 


• 

□ 

'/s 

1 

EH 

V 


■ 

aa 

# 

• 



V 


WMM 

Q C 



• 

• 

• 

• 


m 

Q C 


EH 

UNSHIFT 



• 

• 

• 

• 


• 

•1 

• 

• 

• 

UNSHIFT 


EH 

RUBOUT 



RUBOUT 


mm 


■ 

■ 



I 


L_ 


_ 



DEC 

VALUE UNSHIFT 5 4 3 2 10 SHIFT SUPER SHIFT 


EH 




32 

IF 

8 

”4" 

2 

T 




Dm 

0 








• 


IKSH 


m 

e 







• 



E 


EH 

TJ 







9 

• 

3 

J 


HI 

ELEVATE 






• 



ELEVATE 


cc ec 

EH 

ZUPER SHIFT 






• 

• 

9 

• 



EH 

a 







! A 


UJ UJ 

EH 

CIRCUMFLEX HTCl 






• 

9 

• 

CIRCUN 

SPACE 

IFLEXju^ 


NOTE (UC) ACCENT FOR UPPER CASE CHARAC 

(LC) ACCENT FOR LOWER CASE CHARAC 

0 8 

0 9 

SPACE BAND 









BAND 


“DOT LESS 1 








~9 

1 DOTLESS I 


m 

s 







a 


S 


i i 

_ TT 







LJ 

u 

□ 


1 2 

i 






• 


■ 

r 


1 3 

FIG - 





□ 

□ 


□ 

8 


1 4 

u 





□ 

□ 

<• 


u 


1 5 

s 





u 

u 

U 

u 

7 


1 6 

■mill — 




y 




z 



1 7 

APOS 1 




n 

1 



□ 

i 


1 8 





n 


■ 

U 

■ 

D 


1 9 

HY- 




• 


■ 

u 

□ 

*/• 


2 0 

r 




U 

■ 

u 



R 


m 

7 




U 

■ 

u 


u 

4 

>S 

EH 

j 




LJ 


□ 

□ 


J 


111 

■ 111111 rn^rn 




U 


u 

u 

u 



EH 

n 




□ 

□ 

■ 

■ 

■ 

N_ 


Wl EH 

COMMA 




u 

u 



□ 



B1 

f 


■ 

■ 

u 

• 


• 


F 


EH 

Q L 


■ 

s 

B 

•1 


w 

• 

Q L 


EH 

c 


i 

• 

• 



C 


EH 

D 


z 

z 

• 

• 

• 


□ 

MlIB 


3 0 

R 

■ 

■ 

■ 

u 

u 

• 

w 


K 


3 1 

1 UNIT SPACE 

■ 

■ 

■ 

u 

□ 

9 

9 

w 

1 UNIT SPACE 



t 

■ 

■ 

□ 

■ 





T 


m 

<* 



u 





□ 

5 

CE 

EH 

z 



• 




□ 


Z 


EH 

~T 



□ 

■ 



□ 

□ 

1 


EH 

L 

■ 

■ 

u 



a 



L 


EH 


z 


• 



□ 


m 

' GRAVE (UC) 


1 3 8 

mmnwmwmm 



u 



u 

u 


W 


EH 

s 

L 


u 



u 

u 

u 

2 

? 

EH 

h 

■ 


□ 


□ 



z 

H 


Lai 

DIERISIS (LC) 



• 


• 



u 

EMFIP3] 


WWW 

y 



u 


□ 


u 


Y 


WWW 

& 



□ 


□ 


□ 

□ 

6 

/ 

EH 

p 

m 

■ 

□ 


□ 

□ 



P 


EH 

* 



u 


u 

u 


u 

0 ZERO 


EH 

_s_ 



□ 


□ 

□ 

LJ 


Q 


EH 

h i im hi ii 



u 


u 

u 

LJ 

□ 

'ACUTE(UC) 


EH 

0 


■ 

n 

w 





0 


EH 



■ 

p 

9 




□ 

9 


EH 

b 


■ 

C 1 

u 



u 


B 


5 1 

UR 


■ 

u 

u 



u 

u 

UR 


5 2 

9 



□ 

a 


• 



G 


EH 

• 

» 






• 


□ 

• 

• 


EH 

SHIFT 






•J 

• 

B 

p 

SHIFT 


EH 

L R 






□ 

LR 


EH 

m 





• 



■ 

M 


EH 






Q 

■ 

■ 

□ 



1 58 

X 





y 

11 

u 


X 


Rl 

• □ 





n 

■ 

□ 

□ 

1 

« 

EH 

V 





•1 

• 



V 


[6 1 

Q C 





•1 

n 

■ 

□ 

o 

or 


rwm 

UNSHIFT 




u 

D 

n 

u 

J 

UNSHIFT 


m 

RUBOUT 




□ 

□ 

n 

□ 

□ 



i _ 





_ 




■ 




STD. TYPESETTING CODING 


8 


TAPE 

CHANNEL 

NUMBERS 

CHARACTER 

OR 

FUNCTION 

IQ 

5 

a 

□ 

□ 

□ 

a 

UNSHIFT 

SHIFT 


• 

9 





o 

HDH 


• 




• 

• 

b 




9 


• 

• 



MMBMMM 


• 




• 


d 

D 


• 






• 

E 


• 



• 

• 



f 



9 



• 

• 

9 

G 





• 


• 


H j 



9 


• 




1 


• 

9 



• 


I 

J 


• 

9 


• 

• 



K 



9 




• 

i- 

L 





• 

• 

• 

m 

M 





• 

• 


.. 

N 






• 

• 

© 

o ! 



mm 

SB! 

• 


• 

P 

p 


• 

9 


• 


• 

q 

0 



9 



• 


r 

R 


• 



• 




S 







• 

t 

T 


• 

9 


• 

mm 


HHEHEH 

U 



a 


• 

9 

• 

V 

V 


• 

• 




• 

w 

w 


• 



• 

mm 

n 

> 

X 


• 



• 


• 

y 

Y 


• 





• 

z 

1 

• 

• 



• 

9 

• 

1 

% 

• 

• 

• 




• 

2 

HBEBBB 

• 

* 






3 

HQBl 

• 


• 



9 


4 


• 






• 

5 

H 1 

• 

• 



• 


• 

6 

% 

• 

• 

• 


• 



7 

% 

• 


• 


• 



• 


• 





9 

• 

9 

■■■■■E 

• 


• 


• 


• 

0 

• 

• 




• 

9 

mm 



• 




• 

mm 


COMMA 

COMMA 

• 


• 



9 

9 



• 

• 

• 





s 

1 

KB 

KX 





9 

) 

I I 

• 





9 


APOS. 

QUOTE 

• 

• 




9 


HYPHEN 

-t- 








TAPE 

FEED j 

a 


• 


• 

9 


EN SPACE 

Iks 

• 

• 


• 


9 

EN LEADER 

• 

• 



• 



EM SPACE 

• 




• 


9 

EM LEADER 

• 


• 




9 

VERT. 

RULE J 

• 







THIN SPACE 





• 



SPACE BAR 






9 


CAR RETURN 



• 

. 




ELEVATE 

• 


• 

. 




PAPER 

FEED 


J9 

• 



9 

9 

SHIFT 

(UC) 


9 

• 


• 

9 

9 

UNSHIFT (LC) ! 

• 

• 




mm 

9 

UPPER 

RAIL | 

• 

• 

• 



9 

9 

LOWER RAIL 

• 

• 

a 



9 


STOP (BELL) CODE 

• 

mm 

• 

■1 


9 

9 

RUB 

OUT 

mrm 




• 



ADD 

THIN 

MM 

• 



• 

9 


QUAD 

LEFT | 

■lb 


• 


• 

9 

9 

O 

c 

> 

o 

CENTER 1 

MM 

9 

• 


T 

mm 


QUAD 

RIGHT 

WMM 







ADD 7 

L_ 





_ 


ADO • 


Standard typesetting coding according to Friden. The 
“6” channel becomes the “0” channel in TTS. 


Coding for the V-I-P keyboard comes in two versions: 
left, the U.S. and British, and right, the European ver¬ 
sion. 


60 Composition Input 












































































































































In this section you will find some coding systems used on present day 
devices. Note that most are based on the teletypesetter 6-level code with 
modifications. In any case, the coding structure controls the number of 
keystrokes needed to access any character or command. 

Since machines make mistakes (people are perfect) an additional bit is 
provided in some systems as a doublecheck on proper functioning of the code 
translating device, whether reader or writer. This is the “parity” bit. Using 
one and zero as the basic building blocks of any code produces code chains 
with an even or uneven number of ones. A parity bit is added to make the 
“byte” (a byte is a complete set of bits forming one word) even or odd (as the 
system dictates) and this then allows the system to detect errors and thus 
assure accuracy. 


Character coding for the Mergenthaler V-I-P. 


Mirror image coding is used for certain input systems; 
it “flops” the tape. Thus the character ‘a’ which was 
54 is now 32. 


MIRROR IMAGE CODING 


TAPE 

CHANNEL 

NUMBERS 


8 


CHARACTER 

OR 

FUNCTION 


o 

5 

4 

FEED 

3 

2 

1 

UNSHIFT 

SHIFT 





• 

• 


o 

A 

• 

• 




• 


b 

6 


• 

• 


• 



c 

C 


• 




• 


d 

D 






• 


• 

E 


• 

• 



• 


f 

F 

• 

• 



• 



0 

G 

• 


• 





h 

H 



• 


• 



i 

1 


• 



• 

• 


i 

J 


• 

• 


• 

• 


k 

K 

• 




• 



1 

L 

• 

• 

• 





m 

M 


• 

• 





n 

N 

• 

• 






0 

O 

• 


• 


• 



P 

P 

• 


• 


• 

• 


q 

Q 


• 



• 



r 

R 



• 



• 


s 

S 

• 







t 

T 



• 


• 

• 


u 

U 

• 

• 

• 


• 



V 

V 

• 




• 

• 


w 

w 

• 

• 




• 


X 

X 

• 


• 



• 


y 

Y 

• 





• 


z 

Z 

• 

• 

• 



• 

• 

1 

v. 

• 




• 

• 


2 

V* 






• 

• 

3 

J /« 


• 



• 


• 

4 

Vz 

• 






• 

5 

V. 

• 


• 



• 

• 

6 

V* 



• 


• 

• 

• 

7 

’/. 



• 


• 


• 

8 

Em dash 

• 

• 





• 

9 

& 

• 


• 


• 


• 

0 

0 

• 

• 

• 




• 




• 

• 




• 

COMMA 

COMMA 

• 

• 



• 


• 







• 

• 

• 

$ 

1 

• 





• 

• 

) 

( 


• 





• 

APOS. 

QUOTE 


• 




• 

• 

HYPHEN 

+ 








TAPE 

FEED 


• 

• 


• 


• 

EN SPACE 

• 


• 


• 

• 

• 

EN LEADER 



• 



• - 

• 

EM SPACE 

• 


• 




• 

EM LEADER 

• 




• 


• 

VERT. RULE 







• 

THIN SPACE 


• 






SPACE BAR 

CAR RETURN 





• 



ELEVATE 





• 


• 

LM 

• 

• 



• 

• 


SHIFT (UC) 

• 

• 

• 


• 

• 


UNSHIFT (LC) 

• 

• 




• 

• 

UPPER RAIL 

• 

• 



• 

• 

• 

LOWER RAIL 


• 



• 

• 

• 

STOP (BELL) CODE 

• 

• 

• 


• 

• 

• 

RUB OUT 



• 

•- 



• 

ADD THIN 


• 

• 



• 

• 

QUAD LEFT 

• 

• 

• 

• 

• 


• 

• 

• 

• 

• 

QUAD CENTER 

UM 




• 




ADD 7 

ADD 8 


• •• 

• • • • 
• • • • 

• • • 

• • 
• • • • 
• • • • 

• • • 

• • • 

• • •• 
• • • •• 
• • • 

• •• • 

• •• 

• • • 

• • • • 

• • • •• 

• • • 

• • • 

• • •• 

• •• • • 

• •• 

• •• • • 

• • • • 

• • • 
•••• •• 
• • ••• 
• •• 
• • • • 
► • • 
• • • •• 
• • ••• 
• • • • 
• • • • 
• • • • 
• •• • • 
• • • • 
• • • • • 
• • • 

» • •• 
• • •• 
• ••• 


All 


B12 


A13 


D12 


C15 


D14 


D13 


A12 


B16 


Cl 2 


B11 


B14 


A14 


D15 


A16 


C11 


C14 


B13 


Dll 


DIO 


A15 


C16 


B15 


D16 


A2 


B2 


C2 


D2 


A3 


B3 


C3 


D3 


A4 


B4 


A8 


B8 


D17 


B17 


D21 


Cl 7 


B21 


V-I-P STANDARD FONTS 


UNSHIFT 


o 

u. o 
o 


C13 


CIO 


LAYOUT 


8 


P4S-1 P45-2 P45-3 P34-1 P34-2 


CHARACTER 


PERIOD 


COMMA 


HYPHEN 


$ 


PHOTON 
INSTRUCTIONS 


LENS POS 1 


LENS POS. 2 


LENS POS. 3 


LENS POS. 4 


LENS POS. 5 


LENS POS 6 


LENS POS 7 


LENS POS. 8 


LOWER PAIR 


LINE LENGTH 


SHIFT 


o < 
u- O 
o 


C9 


C4 


A6 


B7 


A10 


D7 


C5 


D8 


D9 


A7 


B1 


C7 


B6 


B9 


A9 


D5 


AT 


C6 


C8 


BIO 


D6 


D4 


A5 


Cl 


B5 


D1 


A18 


B18 


C18 


D18 


A19 


B19 


C19 


D19 


A20 


B20 


A8 


B8 


C20 


A17 


C21 


A21 


D20 


LAYOUT 


P4S-1 


P45-2 


P45-3 


P34-1 


P34-: 


CHARACTER 


M 


W 


Ve 


V* 


% 


Vz 


Va 


% 


% 


Vs 


<A 


% 


Vz 


Vs 




% 


y« 


Vz 


Vz 


% 


% 


Em Minus Em Em I Em 


+ H 


Vz 


% 


Vz 




PERIOD 


COMMA 


% + % / / 


SUPERSHIFT 


A22 


B22 


C22 


D22 


A23 


B23 


C23 


D23 


A24 


B24 


C24 


(A21) 


D24 


LAYOUT 


P45-1 P45-2 P45-3 P34-1 P34 


CHARACTER 


* « < 


% + % 


* * * 


CODE + 
7 BIT 


THIN SPACE 


Min 


□ 


© 


☆ 


@ 


® |TM 


% 


Composition Input 61 






















































































































































































Character and command coding for the Mergenthaler 
Super Quick. 


Two newer code systems are being used with greater frequency in an attempt 
to standardize information transfer. The first is the United States of 
America Standard Code for Information Interchange (USASCII or ASCII) 
and the other is the Extended Binary Coded Decimal Interchange Code 
(EBCIDIC). Each provides a larger array of code possibilities than are 
possible on TTS, the unofficial standard of the industry. These new code 
systems provide a unique code for each character or command rather than 
shift, unshift precedence coding (two frames for one capital letter, for 
instance). 


62 Composition Input 















5. Punched card input systems 


The RACE II System from Warlock Computer Corporation uses punched 
cards rather than perforated paper tape as the means of communication 
between the typesetting keyboard and the phototypesetter. It does not limit 
the capabilities of the keyboard but adds the functions of a data processing 
system, such as: 

Random Access — The random access capability makes it possible to locate 
any text line or group of text lines by simply reading the legend at the top of 
the punched card. Once located, text lines may be changed, deleted, or 
rearranged as desired. 

Editing and Updating — The ability to easily edit or update material such as 
price lists, directories, parts lists, etc., which are subject to periodic reruns 
with changes. 

Manipulation — The ability to manipulate material to be typeset using high 
speed card sorting (unit record) equipment where the size or economics of 
the job do not justify fully computerized storage and manipulation. In this 
way, a variety of reports to be typeset may be produced from one carefully 
organized deck of cards. 

Computer Access — The ability to typeset material which is already stored 
and manipulated in a standard business computer directly without requiring 
special hardware which is not readily available. 

The RACE II Card Input System consists basically of two components, the 
Keyboard Interpeter Unit (KIU) and the Typesetter Interface Unit (TIU). 

Keyboard Interpreter Unit — The Keyboard Interpeter Unit (KIU) 
provides the means for generating punched cards, using a standard 
typesetting keyboard. The unit serves as a link between a typesetting 
keyboard and an IBM Keypunch, Model 026 or 029. 

Typesetter Interface Unit (TIU) — The Typesetter Interface Unit is used to 
transmit data from punched cards to the Phototypesetter system. 

The KIU is connected directly to the punch drive circuit of the typesetting 
keyboard. As the operator of the keyboard prepares the copy, data in the 


Composition Input 63 



code format of the keyboard is routed through the connecting cable to the 
KIU. The KIU circuitry converts the data from the code format of the 
keyboard to Hollerith coding and then drives an IBM Keypunch to enter the 
data on cards. 

The Console of the TIU is interfaced between the tape reader of the 
phototypesetter and the logic circuits of the typesetter. Controls on the 
Console enable the operator to select either paper tape or cards as the input 
medium to the phototypesetter. When in the paper tape mode of operation, 
the signals from the tape reader are routed directly through the output cable 
to the phototypesetter. When in the card mode of operation, signals from the 
card reader are routed to the Converter Cabinet of the TIU where the 
Hollerith formatted signals are converted to TTS code or appropriate 
typesetter code format. The signals are routed through the Console to the 
phototypesetter which reads the data as though it originated on paper tape. 
The Console also makes it possible for the operator to select the column on 
the card where reading will begin, the column where reading will end, and 
the first and last columns within the card which identify a block of 
information to be deleted. 



A standard eighty column punched card. 


Character or row 


Information 

Sections 


An IBM System 3 punched card. 



Tracks or Leve 


Each keystroke at the keyboard will cause one column to be punched on the 
card. The alphabet and the numerals are printed at the top of the card with 
alphabetic characters appearing as capital letters regardless of the shift or 
unshift condition of the keyboard at the time. In addition to the alphabet and 
numerals, the period, comma, hyphen, fraction bar and several special 
characters print out at the top of the card. The special characters are used to 
indicate the most frequently used function codes. This makes it easier to 
determine that a card has been prepared correctly than trying to read the 
punched codes. Some function codes will be punched on the card but will not 
print a symbol at the top of the column. 

64 Composition Input 















The function codes for line length, leading, point size and typeface are 
keyboarded as though the information were being entered on tape. These 
instructions should be followed by a carriage return to insure that the line 
parameters are on a single card which may be replaced, if in error, or 
duplicated, with the duplicate cards inserted in the deck where needed to 
avoid keyboarding the information each time it is required. Normal straight 
matter is keyboarded just as if the information were being placed on paper 
tape, terminating each line with a carriage return. This will automatically 
release the card and register a new one ready for the next line. Under certain 
circumstances, it is possible for the operator to get ahead of the keypunch 
after a carriage return, resulting in the loss of codes on the card. This occurs 
when the carriage on the keyboard completes its return to the left margin 
faster than the keypunch can release one card and register a new one. The 
operator should develop the habit of coordinating the typing operation with 
the keypunch rather than the keyboard. Lines exceeding eighty codes must 
be continued on another card. If the line can be terminated before the end of 
the card, no futher attention is required. If the line must be continued on the 
next card, it is necessary for the operator to depress the REL key of the 
keypunch, wait until the new card has registered and then complete the line, 
ending with a carriage return. 


Certain tabular runs may have columns to be deleted in some galleys and 
added in others. For example: (a) A large number of items that require 
automatic card sorting in accordance with coded categoried i.e. brass, iron, 
plastic, wood parts, etc. (b) Wholesale/retail prices to be produced from the 
same deck, (c) Customer billing information such as in classified 
advertisements. In this case, the Card Length and Field Delete switches are 
used to identify the fields (columns) on the card which are to be typeset. If 
the information to be deleted is at the front of the card, set the Card Length 
Begin switch to identify the last column to be ignored. If the information to 
be deleted is at the end of the card, set the Card Length End switches to 
identify the last card column to be read and the Field Delete End switches to 
identify the last column to be deleted. The system will run faster if 
information to be deleted is at the end of the card rather than the beginning. 
This is due to the fact that the phototypesetter begins flashing a line 
following carriage return while the reader ejects the card, inserts a new card, 
and then waits for the “START” signal from the phototypesetter before 
reading the new card. It is more efficient to allorbefore reading the new card. 
It is more efficient to allow the card reader to pass over deleted information 
while the phototypesetter is flashing, than to force the phototypesetter to 
wait during this period. 


In the normal mode of operation, keystrokes are duplicated on cards column 
by column. That is, each keystroke produces a corresponding punched code 
in the card. The alphabet and the numerals are printed at the top of the card 
and since the keypunch has only one “case”, the legend at the top of the card 
will always appear as capital letters or numbers regardless of the “shift” or 
“unshift” condition of the keyboard. 


Composition Input 65 



In addition to the alphabetical characters and numerals, periods, commas, 
hyphens, fraction bars and several additional special characters are also 
printed on the top of the card. The special characters are used to indicate the 
most frequently used function codes which makes it easier to determine if a 
card has been correctly punched. 

It is a matter of convenience whether or not function codes are placed on the 
same card with the text matter or are placed on separate cards. Placing the 
function codes on the same cards with the text will sometimes reduce the size 
of a deck significantly as well as eliminate the possibility of misplacing of 
these codes. On the other hand, if function codes are placed on separate 
cards, it will be easier to change formatting of instructions such as leading 
and font, in case this should be required after initial keyboarding is 
completed i.e. increasing or decreasing the leading point in order to achieve a 
better copy fit. 

The flexibility of the keypunch and the RACE II Card Input System as well 
as the variety of keyboards which can be used with this system make it 
impossible to anticipate all possible combinations of tab requirements. The 
following paragraphs, therefore, describe a few typical examples in order to 
illustrate the basic principles of tabular work. 

Examine the list for longest part number, longest nomenclature, or longest 
retail and wholesale prices, including the dollar sign. Keyboard the 
composite “longest line” using fixed spaces between part number and 
nomenclature, and between retail and wholesale prices, and use J-spaces 
between nomenclature and retail price. A test setting of the text line can be 
made to determine the final line length to be used in the typesetter or else, 
the line length can be calculated from a character count and/or the character 
width tables available in the appropriate typesetter Manual. 

From this card, a program card for the keypunch is made up which causes 
the RACE II Card Input System to perform repetitive operations 
automatically. 


The sequence of subsequent operations is as follows: 

(1) The keypunch registers a card, punches the lower case code and then 
stops. 

(2) The operator keyboards the part number and depresses the SKIP key. 

(3) The keypunch punches the tab codes, shift code and “Quad Left” code 
for the nomenclature column, and then stops. 

(4) The operator keyboards the nomenclature, shifting and unshifting as 
necessary, and depresses the SKIP key. 

(5) The keypunch skips over the remaining columns of the nomenclature 
field, and then punches in the tab codes and a “Quad Right” code for the 
retail price column and stops. 


66 Composition Input 



(6) The operator keyboards the dollar sign and/or fixed spaces plus the 
retail price, as required, and depresses the SKIP key. 

(7) The keypunch punches in the tab codes, the “Quad Right” code for the 
wholesale column, and stops. 

(8) The operator keyboards the wholesale price in the same manner as the 
retail price, described in (6) above, and depresses the SKIP key. 

(9) The keypunch releases the card, registers a new card, punches the unshift 
code, and stops, ready for the next line. 


HOLLERITH CODE 


TTS CODE 

CHARACTER 


HOLLERITH CODE 


CHARACTER 


TTS CODE 


9 

8 

7 

6 

5 

4 

3 

2 

1 

0 

11 12 

! 

I 

s 

i] 


3 


I] 

I] 

3 














0 



IT 


7T 

• 

• 











— 





1 



• 

• 

• 

• 


• 









— 






2 



• 



• 

• 

• 








— 







3 






• 


• 







— 








4 




• 


• 

• 







— 









5 



• 



• 







— 










6 



• 


• 

• 


• 




— 











7 





• 

• 

• 

• 



— 












8 





• 

• 

• 



— 













9 



• 

• 


• 












— 



— 


A 






• 

• 

• 









— 




— 


8 



• 

• 




• 













— 


C 




• 

• 


• 








— 






— 


D 




• 




• 






— 







— 


E 








• 













— 


F 




• 

• 



• 




— 









— 


G 



• 

• 


• 

• 




— 










— 


H 



• 


• 

• 




—i 













1 





• 

• 

• 











— 


— 



J 




• 


• 

• 

• 









— 



— 



K 




• 

• 

• 

• 

• 








— 




— 



L 



• 




• 








— 





— 



M 



• 

• 

• 

• 








— 






■■ 



N 




• 

• 

• 







— 







— 



0 



• 

• 


• 






— 








— 



P 



• 


• 

• 

• 




— 









— 



Q 



• 


• 

• 

• 

• 


— 










— 



R 




• 


• 

• 










— 


—i 




S 





• 

• 


• 








— 



— 




T 



• 



• 









— 




— 




U 





• 

• 

• 

• 






I— 





— 




V 



• 

• 

• 

• 

• 






— 






— 




W 



• 



• 

• 

• 




— 







— 




X 



• 

• 

• 

• 


• 



— 


1 





— 




Y 



• 


• 

• 


• 


— 



j 





— 




z 



• 



• 


• 





i 



— 


■■ 




/ 






• 

• 






t 

_L 
















A conversion chart of Hollerith (punched card codes) 
and TTS coding. 


□ 

□ 

□ 

□ 

3 

a 

□ 

□1 

a 

□1 

ED 

ED 


[ 

3 

3 

3 

3 

3 

3 

3 

3 

3 

l 

1 

1 

1 

1 

1 

1 

II 

i 

II 

1 

9 


CARR. RETURN 






• 




■ 

B 

■ 

■ 

1 

■ 

■ 

a 

■ 



— 


o/ 

/o 



• 



• 

• 


• 

□ 

H 


■ 

■ 

■ 

3 





—i 


PERIOD 



• 

• 

• 

• 



• 

■1 

□1 




— 






— 


LINE LENGTH 


• 


• 

• 

• 




■1 

□1 



— 




■ 

■ 

■I 

3 


EM SPACE 





• 

• 


• 


□ 

— 


□i 

■1 

■ 






— 


EN SPACE 




• 

• 

• 

• 




□ 

□1 









— 


QUOTE 




• 


• 





3 






— 



— 



1 






• 

• 




3 




H 

El 

■1 


■ 

□ 

■ 


ADD LEAD 


• 




• 

• 




3 



— 






— 



UM 




• 

• 

• 

• 




□ 


— 







— 



LM 






• 

• 




3 

□ 








— 



( 



• 



• 


• 


■ 

□ 

■ 

■ 

■ 


B 



— 




COMMA 




• 

• 

• 




■ 

□ 


■ 

■ 

3 




— 




BELL 




• 


• 

• 

• 


i 

i_ 

— 


■ 

31 

■ 


■ 

■ 

□ 

■ 

■ 


BLR 





• 

• 




L 

— 


— 

1 


■ 

■ 


-» 




EM LDR 



• 


• 

• 




■ 

□ 

□ 







□ 

■ 



EN LDR 



• 


• 

• 

• 

• 












— 



HYPHEN 




• 




• 



— 


■ 

■ 

■ 

E 

■ 






UPPER CASE 



• 

• 



• 

• 



— 


■ 

■ 

a 


■ 






LOWER CASE 



• 

• 

• 


• 

• 


m 

□ 

■ 

■ 

- 









UPPER ROW 



• 

• 


• 


• 

• 

i 

_ 

1 


2 

m 









. LOWER ROW 



• 

• 


• 

• 

• 

• 









□ 

E 


— 


QUAD LEFT 




• 

• 

• 


• 

• 








— 


E 

■ 

E 


QUAD CENTER 



• 

• 

• 

• 

• 


• 







□ 

■ 

■ 

— 


— 


QUAD RIGHT 


• 




• 

• 


• 





it 

Q 




— 


— 


SEMI COLON 



• 

• 


• 

• 


• 





— 





— 


— 


THIN SPACE 






• 



• 




13 

■ 

■ 

■ 

m 


— 


— 


IU SPACE 


• 




• 

• 





— 


■ 

■ 

1 



— 


— 


EOL LEAD 


• 




• 

• 

• 



— 

— 











SPACE BAND 





• 

• 




* 

m 

■ 




IE 

m 

■ 

■ 

IE 

IE 

■ 

LENS ONE 


• 


• 


• 

• 

• 

• 






—i 





IE 

IE 

■ 

LENS TWO 


• 


• 


• 

• 

• 


■ 

■ 

■ 

II 

IE 

■ 

II 

m 



— 

— 


LENS THREE 


• 


• 


• 

• 





■ 

IE 

■ 

■ 

II 




— 

— 


LENS FOUR 


• 


• 


• 

• 


• 

■ 

■ 

— 



■ 

II 

m 


m 

IE 

IE 

■ 

LENS FIVE 


• 


• 


• 



• 

■ 

IE 

■ 








IE 

IE 

1 

LENS SIX 


• 


• 


• 




— 

■ 

■ 

■ 


■ 

■ 

■ 


n 


E 


LENS SEVEN 


• 


• 


• 


• 








L_ 

_ 


IE 

IE 

II 

11 

LENS EIGHT 


• 


• 


• 


• 

• 

■ 

IE 

m 

H 

li 

i 

■ 

IE 

ii 

IE 

11 



TAPE FEED 






• 




L 

— 




wm 



ii 

II 

IE 



TAB (713-5) 


• 



• 

• 


• 


■ 

1C 

m 

II 

II 

IE 

■ 

■ 

■ 

II 

— 



TAB (713-10) 


• 

• 



• 


• 







■ 

■ 

11 

r 

E 




STOP 


• 




• 





Composition Input 67 


















100111 BRAHMS - Sonata In F Minor For Clarinet And Piano/$13. 
Sonata In E-Flat Major For Clarinet and Piano - 
Harold Wright, Clarinet; Harris Goldsmith, Piano 


The fixed information in this example is much greater than the variable 
information. Furthermore, for the purposes of ths example, it is also 
assumed that the variable information changes frequently. Examine the copy 
for the longest part number and the longest price. Either a test setting can be 
made, or else the width can be calculated from a character count and/or 
character width table. The space remaining can be devoted to the 
nomenclature and description columns. Since the price changes frequently 
and only represents a small part of the total text, it is best to place it on a 
card by itself. Keyboarding would proceed as follows: 

(1) With the regular program card on the keypunch and the starwheels 
down, keyboard part number, “TAB”, the first line of text matter to 
justification, “TAB” and “Quad Right” codes, and depress the “REL” key. 

(2) Keypunch registers a new card. 

(3) Operator keyboards the price and a “Carriage Return”. 

(4) Keypunch registers a new card. 

(5) Operator keyboards “TAB”, the second line of text matter to 
justification, and a “Carriage Return”. 

(6) Keypunch registers a new card. Additional lines of text matter are 
keyboarded as in (5) above until the text is completed. 

(7) Start the second item as in (1) above. 

Price revisions can be keyboarded most efficiently in one continuous 
operation and be inserted into the card deck in a second operation. This will 
reduce greatly the possibility of human error. The use of colored cards can 
be a help in locating price cards or in keeping track of new versus old prices. 


The following is an explanation of the terms used in the RACE II Card 
Input System. 

Column 

The RACE II Card Input System uses standard 80 column cards. A column 
has twelve possible punch positions, sometimes referred to as “twelve 
levels”. Each column is equivalent to one code frame on perforated paper 
tapes. Columns are numbered 1 through 80 with column numbers normally 


68 Composition Input 



printed on the face of the card. In conversation be careful to distinguish 
between a “card column” and a “column of tabular copy”. 

DUP 

This stands for “Duplicate”. The use of this keypunch key permits 
duplication of information from one card to another. However, it does NOT 
duplicate information from the program card. 

Field 

A group of columns on a punched card may be referred to as a “Field.” For 
example, a group of columns on a punched card which contains a price might 
be referred to as the “Price Field”. Similarly there may be a “Date Field”, a 
“Name Field”, etc. 


Format Card 

This term refers to a card which contains format instructions for the 
typesetter. Do not confuse it with a Program Card used on the Keypunch. 

Line Card 

A card which contains nothing but a carriage return (line space) code is 
called a “line card”. It produces an amount of leading which is equal to the 
primary leading programmed into the typesetter. 

Program Card 

A program card is used on the program drum of the keypunch to control the 
automatic functions of the keypunch. No information is transferred from the 
program card to the card currently being punched. The program card 
excercises control over which columns of information are to be duplicated 
from the card in the read station to the card in the punch station, or which 
columns are to be skipped entirely. 

Program Drum 

This is the drum in the keypunch which holds the program card. When the 
starwheels are down, the program card controls the automatic functions of 
the keypunch. When the starwheels are in the up position, control of the 
keypunch is performed manually. 

Rel 

This key on the keypunch causes the card in the punch station to be released 
and a new card to be registered. 

Row 

The cards used by the RACE II Card Input System have twelve “rows”. The 
“twelve” row is located at the top of the card, immediately below the printed 


Composition Input 69 



legend. The next row then is 
“four”, “five”, “six”, “seven”, 


“eleven”, “zero”, “one”, “two”, “three”, 
“eight”, “nine”, in that order down the card. 


Run-Out Card 


A card with sufficient “add-leads” to advance the paper in the magazine 
after a run on the typesetter is used in place of the manual leading. 

Skip 

Depressing this key on the keypunch causes the card to advance rapidly to 
the next field as defined by the program card. With the program card used 
for normal text, the card will skip the remainder of the card, and a new card 
will be registered in the punch station. 

Stop Card 

This is a card with a “stop” code plus a “tape feed” (blank frame or column) 
code which is placed at the end of a deck of cards to be run. If it is at the end 
of a take, it should be preceded by a “run-out” card. 




The Mergenthaler V-I-P, which stands for Variable 
Input Phototypesetter. It incorporates its own 
minicomputer. 


70 Composition Input 













6. Video display terminal (VDT) systems 


There are over 200 video display terminals on the market. Some are the 
alphanumeric type and show numbers, letters and special symbols on the 
screen. Others are “graphic” terminals with the ability to draw lines on the 
screen and thereby display representations of three-dimensional objects. 

The video display terminals used in typesetting today are alphanumeric, 
varying from simple models which can display only a few hundred cap 
characters to those which may indicate different “typefaces” on the screen. 

A video display terminal is a special kind of television. An electron gun 
projects a stream of electrons onto a phosphor coating on the face of the 
tube, causing the tube to glow. The color of the image varies with the kind of 
coating inside or the tinted screen which is over the front of the tube. 

The phosphor coating is excited to a state of fluorescence that has a degree 
of persistence (it takes a time for the image to fade). If the phosphor fades 
quickly the image fades likewise. Because the image projected on the face of 
the tube begins to fade it must be refreshed or displayed over and over again 
for the image to appear stable on the face of the tube. This means that the 
information to be displayed must be held in some sort of buffer memory and 
recycled many times a second to achieve a reasonably flicker free image. 

Video terminals may be configured in a wide variety of different ways. There 
is the totally self-contained editing terminal which does not need a computer. 
This unit reads text in from paper tape, displays it on the screen to be 
modified and then repunches the corrected version as a clean tape. 

There is also the slave unit with virtually no logic of its own. This unit must 
be wired directly to a computer. In between these units is the stand-alone 
computer terminal connected to a computer to accept or deliver hunks of 
text, but with its own logic for storage of the text to be displayed on the 
screen and for making changes in this text. Lastly, there is the cluster 
terminal which shares logic among several separate stations. 

Characters are formed on the face of a VDT tube in the following ways: 

1. Characters are created out of a series of dots. Most are designed to form a 
rectangle five dots wide by seven dots high (a 5 x 7 dot matrix). It is difficult 
to design legible lowercase letters within these constraints and some VDT’s 


Composition Input 71 



have gone to a 7 x 9 or greater matrix. A dot matrix pattern is well suited to 
a television tube with its fixed raster scan. This can, in effect, make any TV a 
terminal. 

2. Characters are created out of a series of line segments. This is called 
vector generated characters and is less commonly used. 

3. Another approach is to “paint” the characters with a series of horizontal 
or vertical strokes (raster scan), in a manner similar to that used on most 
CRT typesetters — although, of course, with much lower resolution. 

The number of characters which may be displayed on the tube at one time is 
a function of the resolution of the tube. The more addressable points there 
are on a tube, and the fewer addressable points required to describe each 
character, the more characters will fit on the screen at one time. The size of 
the tube itself is of relatively little consequence in determining how much 
information will fit on the screen. You get the same picture information on a 
7” T.V. tube as you do on a 21” tube. You may prefer the 21” tube because 
it is larger, but all you are really seeing is an enlarged and more legible 
version of the same picture. 

Most of the VDTs which are commercially available are not aimed 
specifically at the typesetting market and have limited character sets — 
often about 64 separate symbols, which means upper case characters only 
and very few special symbols. Increasingly, greater numbers of upper and 
lower case alphanumeric terminals are coming on the market. Most of these 
will display over 96 separate symbols. Some will display as many as 128. 


There are three basic elements common to VDT units: a) Refresh logic, b) 
Character generation, and c) Editing electronics. As characters enter the 
VDT via keyboard, paper tape or by direct cable, they are stored in the 
refresh logic. From here the characters are projected on the face of the 
cathode ray tube sixty times a second. This cycle is necessary so that the 
image on the screen appears stable. The shape and characteristics of each 
character are produced by character generation circuitry. 


On most units, the character location on the screen at which the on-going 
operation takes place is defined by a “cursor.” This may be an underline 
dash or an entire rectangle, often having a “blinking” appearance for ease in 
location on a screen full of characters. The cursor is the key to most VDT 
editing. It is positioned by means of control keys that move it up, down, left 
and right. Some VDTs also have a “home” key to bring the cursor to the 
first location in the left hand corner of the screen. Additonal commands may 
move it to the end of a paragraph or to other locations. It is also used to 
delineate the character, word, line or paragraph under editing scrutiny. 

VDTs incorporate standard typewriter (QWERTY) sets of keys with two 
extra sets: one for editing and one for typesetting functions. 

Here is a list of the editing and function controls and what they do for one 
specific VDT. This list is by no means inclusive, and presented only to orient 
you to the kinds of command keys available on VDTs. 


72 Composition Input 



CONTROL or 
INDICA TOR 

FUNCTION 

FUNCTION 

SUPPRESS 

Switch 

Toggle switch. In up (Suppress) position inhibits display 
of function symbols on screen, but does not delete them 
from memory. In down (Normal) position, allows func¬ 
tion symbols to be displayed. 

JUST/UNJUST 

Switch 

Toggle switch. In JUST position, causes Elevate code 
read from tape or keyboard to move cursor to new line. 
In UNJUST position, Elevate codes are displayed and 
punched, but are not executed. 

EDIT/PERF 

Switch 

Toggle switch. In EDIT position, all functions are ac¬ 
tive. In PERF position, characters typed on the keyboard 
are punched one by one. The screen and tape reader are 
disabled. 

BRIGHTNESS 

Potentiometer 

Controls brightness of characters displayed on screen. 
Rotate clockwise to increase brightness, counterclock¬ 
wise to decrease brightness. 

TAPE LOW 

Lamp 

Lamps lights when tape-out switch on punch tension 
plate opens, indicating that only a few inches of blank 
tape remain. Punch is disabled, but unpunched text 
remains intact in punch buffer. By setting EDIT/PERF 
switch to PERF, leader codes may be punched before 
loading new spool of tape. 

K/BSHIFT 

Lamp 

Lights whenever machine enters shifted mode; goes out 
when machine returns to unshifted mode. 

RESET 

Switch 

Pushbutton switch. When pressed, erases entire screen, 
drives cursor to Home position, sets unshifted mode, 
and initializes reader and punch interface. 

QL Key 

Displays Quad Left symbol and drives cursor to start 
of new line. 

QC Key 

Displays Quad Center symbol and drives cursor to start 
of new line. 

QR Key 

Displays Quad Right symbol and drives cursor to start 
of new line. 

UR Key 

Displays Upper Rail symbol. 


Composition Input 7 3 



RUBOUT Key Displays Rubout symbol. 


LR Key 

Displays Lower Rail symbol. 

RET Key 

Displays Return symbol. 

BELL Key 

Displays two separate symbols for shifted and unshifted 
modes. 


Alternate action key; if machine is in unshifted mode, 
SHIFT LOCK SHIFT LOCK sets shifted mode and lights K/B SHIFT 
Key lamp; if machine is in shifted mode, SHIFT LOCK sets 

unshifted mode and extinguishes K/B SHIFT lamp. 


ELEV Key Displays Elevate symbol. In JUST mode only, drives 

cursor to start of new line. 


TAPE FEED Key Displays Tape Feed symbol; has no immediate effect 

on punch. 


HOME 

Drives cursor to Home position (top left corner of 
(screen). 

(Cursor Up) 

Key 

Drives cursor up one line for each pressure. Inhibited if 
cursor is already in top line. 

DEE START 

Key 

When pressed, current cursor position is defined as start 
of punch or delete block operation. Cursor must then 
be moved to indicate end of operation. 

(Cursor Left) 

Key 

Drives cursor one character position left for each pres¬ 
sure. Inhibited if cursor is already in leftmost position 
of line or column. 

NEW LINE Key 

Drives cursor to start of next line. 


(Cursor Right) Drives cursor one character position right for each pres- 
Key sure. Inhibited when cursor reaches end of screen. 


(Cursor Down) Drives cursor down one line for each pressure. Inhibited 
Key when cursor reaches bottom line. 


Starts tape reader. Reader stops automatically after 
READ TAPE approximately 1600 characters have been read and dis- 
Key played at 50 cps. rate. After the halt, reading may be 

continued by pressing READ TAPE key for each char¬ 
acter or by holding key down to read at Repeat rate of 
15 cps. 


Halts automatic reading at any time. In automatic read 
STOP READ area of screen, may be held down while READ TAPE 
Key key is pressed to read single characters. 

-^—i--- 

74 Composition Input 



PUNCH TAPE 
KEY 

When pressed, a block of text is transferred to the punch 
buffer and punched. The start of the block is defined by 
the cursor position at the time the DEF START key is 
pressed; if no start position is defined, punching starts 
from the Home position. The end of the block is the 
character to the left of the cursor position at the time 
PUNCH TAPE is pressed. 

INS CHAR 

Key 

The character at the cursor position and all characters 
to the right of it are moved one position right; a null is 
inserted at the cursor position. The function is inhibited 
if the shift would move any character except a space into 
the last position of the line. 

DEL CHAR 

Key 

The character at the cursor position and all characters 
to the right of the cursor are moved one position left. 
A null is inserted in the last position of the line. 

OPEN Key 

All characters from the cursor position to the end of 
text are recopied starting at the end of screen and build¬ 
ing up. 

CLOSE Key 

Material at the bottom of the screen is recopied, starting 
at the cursor position. This closes any gap left after in¬ 
sertion of new material; the wraparound feature prevents 
the breaking of words at line endings. 

DEL BLOCK 
Key 

Erases all characters between a previously defined Start 
(or the Home position if no Start was defined) and the 
current cursor position. The gap is closed. 

CLR Key 

Erases all characters from the current cursor position 
to end of screen. Nulls are inserted in the erased 
positions. 



A full view of the CorRecTerm keyboard. There are 
three basic, if undemarcated areas: typewriter charac¬ 
ter set, typesetter function set and display control set. 


Composition Input 75 






Editing 


When a character or group of characters is presented on the screen, certain 
editing functions are possible, more or less dependent upon the particular 
VDT. Most errors involve single characters, and these are corrected by 
positioning the cursor at the character position involved and depressing the 
correct character key. This erases the incorrect character and replaces it. 
The technique is called “writing over.” On some units it may be necessary to 
first strike a DELETE key to erase the incorrect character, hit an INSERT 
key and then the correct character. Writing over is much simpler. 

Characters are either inserted, deleted or changed. So are lines and 
paragraphs. Thus editing keys may be provided to perform each of these 
functions. In all these operations the cursor is used to define the data to be 
operated on. Positioned at the beginning of the word or paragraph, new data 
may be inserted or removed. When deleted, most units automatically close 
up the resultant space by moving all characters to the right of the “hole” left 
to fill it up. Insertion of characters moves all data right to make room. An 
important function of VDTs is the ability to “wraparound”; that is, take 
words that will not fit on a line and move them to the beginning of the next 
line and so on until the end is reached. 

The operator has now input all data, modified it through changes, additions 
and deletions and is now ready to be output in the form of tape or direct 
impulses to a photo unit. 

Stand-alone units 

+ 

Take one VDT “tube,” put a reader on one side and a punch on the otherand 
you have a stand-alone unit. You can read in tapes produced on other 
keyboards, or type in information from the keyboard. The resulting data 
may be reviewed and edited and then output via the punch to produce a new 
tape. 

Expanded systems 

If a great deal of data will be edited, a magnetic disc or drum or tape cassette 
may be used to store all characters before and after their screen debut. 
Normally a “controller” which is often a mini-computer directs character 
traffic between the storage medium and the VDT. 

Expanded VDT systems may also involve dependence upon larger general- 
purpose computers. Here much greater capability is available for massive 
amounts of editing, storage and retrieval. 

VDTs began as what data processing folks call “data windows.” The first 
units were replacements, not for keypunches or typewriters, but for 
teletypes. Teletypes are widely used for computer input. Systems progressed 
and were tied into computers for reservations, credit checking and inventory 
control. The late, great company called Viatron attempted to create a unit 
that would pre-process data prior to computer input. The year Viatron bit 
the dust over forty-five VDTs were exhibited at the Spring Joint Computer 
Conference. Many more were certainly in the wings. 

76 Composition Input 



It was a little under three years ago when the first VDTs were shown to 
anxiously - waiting newspapers at the ANPA/RI exhibit in Chicago (1969). 
It has taken some time but we are just getting started. 


The Hendrix 5700 video display. 



Some important points: 

The size of the screen is important. Both for the amount of data to be 
handled at one gulp and to the clarity and size of characters. VDTs are a 
visual medium, and like television, can contribute to eye fatigue. One of the 
larger units today has an 18 inch (diagonal) screen that displays 30 lines of 
90 characters. About 2,000 characters is the average. 

Erasing the screen gets rid of material which is no longer needed. Used 
carelessly, it will also get rid of material which is needed. Character, Cursor- 
to-end-of-line, and Line Delete functions handle short erasures. 

Scrolling may vary from simple movement of displayed material upward, off 
the top of the screen both upward and downward movement. 

Scrolling depends largely upon the relationship between the VDT’s refresh 
logic and its screen complement. When memory capacity equals the number 
of characters which can be displayed, memory content is exactly the same as 
the display. Blanks are stored in memory for all unused positions on the 
screen. More sophisticated units are designed so that only the characters 
which are displayed take up space in the refresh memory. 


Composition Input 77 




Here is the keyboard arrangement of one VDT system presently on the 
market. Note the extensive editing capabilities available. 



78 Composition Input 


s c ur Sort Keys C&*?4r) 


















































1. CLEAR MEMORY (MAIN) 
COPY MAIN INTO AUXILIARY 


25. ADD THIN SPACE 
EM SPACE 


2. COPY AUXILIARY INTO MAIN 
SWAP MAIN AND AUXILIARY 


26. QUAD RIGHT 
EN SPACE 


3. LEFT MARGIN CONTROL 27. RETURN CURSOR (entry) 


4. RIGHT MARGIN CONTROL 28. CURSOR RETURN (editing) 


5. DISPLAY MODE SELECTOR 

6. RESET SYSTEM LOGIC 


29. CHANGE CASE 

INSERT MODE SELECTOR 


30. CHANGE BOLD 

7. INTERLACE ON/OFF BOLD SELECTOR 

HIDE CONTROL CODES ON/OFF 

31. CHANGE DIM 

8. REPEAT DIM SELECTOR 


9. VERTICAL RULE 

10. BELL CODE 

11. TAPE FEED 


32. LINE INSERT 
CHARACTER REMOVE 

33. LINE REMOVE 
WORD REMOVE 


12. UPPER RAIL 

13. PAPER FEED 

14. THIN SPACE 

15. HERE IS MULTIPLE CODES 


34. PARAGRAPH REMOVE 
SENTENCE REMOVE 

35. SET/SKIP 

36. REMOVE TO CURSOR 
CURSOR UP 


16. ELEVATE 


37. SET/TAB 


17. EN LEADER 


38. CURSOR LEFT 


18. EM LEADER 

19. RETURN 

20. SHIFT 

21. UNSHIFT 

22. QUAD LEFT 

23. LOWER RAIL 

24. QUAD CENTER 
RUBOUT 


39. END OF TEXT (ETX) 44. STRIP OPTION SELECTOR 

HOME CURSOR 

45. CURSOR I/O SELECTOR 

40. CURSOR EXTREME RIGHT MEMORY OUTPUT SELECTOR 
CURSOR STEP RIGHT 

46. INPUT DEVICE SELECTOR 

41. CLEAR LINE INPUT START/STOP 

ROLL DOWN 

47. OUTPUT DEVICE SELECTOR 

42. CLEAR REST OF MEMORY OUTPUT START/STOP 
CURSOR DOWN 

48. ADD OPTION SELECTOR 

43. HOME MEMORY 

ROLLUP 49. FEED TAPE SPACING 


Composition Input 79 



Some VDT’s do not display some typographical characters and commands. 
Here are some of those symbols displayed by one VDT system. 


Quad Right (QR) 


Quad Center (QC) 

4-> 

Quad Left (QL) 

«— 

Upper Rail (UR) 

t 

Lower Rail (LR) 


Em Space (EM) 

T\ 

En Space (EN) 

□ 

Em Leader (EM LDR) 

• • 

En Leader (EN LDR) 

1 

• 

Vertical Rule (VR) 

1 

Tape Feed (TAPE) 

t f 

Paper Feed (PAPR FEED) 

P F 

Return (RWT) 

3 

Elevate (ELEV) 

•u 

Shift (SHFT) 

T 

Unshift (UNS) 

X 

Thin Space (THIN) 

II 

Add Thin Space (ADT) 

III 

Bell (BELL) 

A 

Rub Out (RUB) 

• • • 

• • • 

• • • 

• • • 


The Hendrix 5200 video display. 



80 Composition Input 





Here are the control keys of the Mergenthaler 
CorRecTerm video display terminal. The cluster of 
keys at right controls editing functions; that at left 
controls cursor position; and that above controls in¬ 
put and output. 



Control of a VDT 



The cursor, displayed by alternating the background of a character between 
black on white (BOW) and white on black (WOB), indicates the position on 
the screen at which a function is to be performed. It blinks to take the 
operator’s attention to that point on the screen. A character entered from the 
keyboard appears on the screen exactly where the cursor is currently located, 
then the cursor automatically moves to the next character position. 

Return Designed to be used for test entry and its function depends on which 
mode is operating, as follows: 

FORMAT MODE — RETURN is used to return the cursor to the 
beginning of the next line on the display. 

JUST or NW MODE — RETURN inserts customer specified line delimiter 
at the end of a line and moves the cursor to the beginning of the next line. 

UNJUST MODE — RETURN inserts customer specified paragraph 
delimiter at the end of a paragraph and moves the cursor to the beginning of 
the next line. In this mode, all cursor-returns at the end of lines are done 
automatically. 

CR (CURSOR RETURN) — M oves the cursor to the beginning of the next 
line. CR was designed to be used for text editing and to be independent of the 
mode of operation. 


HOME —- Moves the cursor to the upper left-hand corner position on the 
screen called “Home.” 


By depressing any one of the arrow keys, the cursor is moved one position at 
a time in the direction of the arrow and will repeat in that direction when 
held down. 


END OF TEXT — Moves the cursor to the exact end of text which is 
defined as the point where the operator can start entering new characters in 
order to append previously entered text. If the end of text position is not on 
screen, ROLL-UP’S will be automatically generated to bring it into view 
where the ETX function can be performed. 

RIGHT — Moves the cursor to the right side of the line the cursor is located 
in. 


Composition Input 81 





Editing 


Many advanced editing controls are often provided so that the operator can 
rapidly and effectively edit text. For the convenience of the operator and to 
help speed up the editing process, a WORD, SENTENCE or 
PARAGRAPH can be removed in their entirety with just a few key strokes, 
in addition to removing single characters at a time. 

OVERSTRIKE is the normal keyboard mode of operation where each 
character entered from the keyboard replaces the character in the location of 
the cursor. The cursor will step right and no other character on the screen 
will be disturbed. 

The operator should select the INSERT mode (INSERT light ON) when he 
desires to insert a missing character between two characters. A character 
entered in this mode is inserted at the position of the cursor, pushing the 
character previously there to the right one position. 

The keyboard is automatically switched from the INSERT mode to the 

OVERSTRIKE mode when the cursor is moved by any of the cursor control 
keys. 

CHAR (CHARACTER) removes the character in the cursor position, 
closing text up. Keeping CHAR depressed continuously removes characters 
by pulling text in from the remainder of the line. 

CCASE (CHANGE CASE) changes the case of the character under the 
cursor from upper case (A,B,C) to lower case (a,b,c) or vice versa. Especially 
helpful when an input tape has missing unshift or shift codes. 

BOLD displays characters as they are entered from the keyboard as black on 
white background instead of the normal white on black. BOLD keyboard 
entry (BOLD light ON), is useful when an operator wishes to distinguish 
bold face characters from the remaining text. This saves the operator from 
having to enter upper and lower rail codes. 

CBOLD (CHANGE BOLD) reverses the display of the character under the 
cursor from NORMAL to BOLD or from BOLD to NORMAL, depending 
on the initial state of the character. Useful for going back over text and 
changing characters to or from BOLD face. 

DIM displays characters entered from the keyboard as either: gray 
characters on black background for DIM/NORMAL ENTRY (DIM light 
ON only); or the reverse, displayed as black characters on a gray 
background for both DIM/BOLD entry (DIM and BOLD lights ON). 
Useful to distinguish italic or other type characters from the rest of the 
displayed text. 

CDIM (CHANGE DIM) reverses the display of the character under the 
cursor from either: NORMAL to DIM/NORMAL or DIM/NORMAL to 
NORMAL; or from BOLD to DIM/BOLD or DIM/BOLD to BOLD. 
Useful for going back over text and changing NORMAL and BOLD 
characters to or from their DIM configurations. 


82 Composition Input 



WORD removes the word in which the cursor is located, closing text up. 


SENT (SENTENCE) removes the sentence containing the cursor, closing 
text up. 

PARA (PARAGRAPH) removes the paragraph containing the cursor, 
closing text up. 

LIN (LINE INSERT) moves all text from the line the cursor is located in 
down one line, leaving a blank line. 


LRM (LINE REMOVE) removes the line containing the cursor, moving the 
remaining text on the screen up one line to fill the space. 

CLL (CLEARLINE) erases the line containing the cursor, leaving a blank 
line in its place. 


Rolls text on the screen up one line at a time “forward” through main 
memory, until it stops at the bottom of memory. RU can be Roll up, Roll 
Down the beginning of the story. Rolls text on the screen down one line at a 
time “backward” through main memory, until it stops at the top of memory. 


«Ihe Hergenthaler Coi 
systei for proofing, cc 
editing text ~ prior 
typesetting systeis.I 
Tie fferoenthaler C r 

m <>ri 

'^sistifiQ of > , v 



level paper tape 
and a combination CRT 
p iput (TTS) keyboard. 

* nd aesthetically, these 

ideally suitted to 
or plant.[ 
er« serves these 
typesetting features:! 

5 fast, one-tan tethtd of 
ig and tape terging of 
ections.1 


Remits easy insertion of fwction 
or paraaeter codes to coaputer 
input tape.! 


..reader 
•».e that 
■*nd 




The Mergenthaler CorRecTerm. The unit at right 
reads in tapes and perforates new ones. 


Composition Input 83 







Operation 

Keystrokes 
per Hour 

Ems 
Approx, 
per Hour 

-Equiv. 

Typewritten 

Words/Min. 

1. Manual Linecaster 

7,000 

3,500 

23 

2. Monotype Keyboard 

10,000 

5,000 

32 

3. TeleTypesetter Keyboard 

10,400 

5,200 

33 

(Justified Lines) 

4. Typewriter Keyboard 

18,000 

' 9,000 

59 

(Non-justified Tape) 

5. Typewriter 

23,800 

11,900 

78 


One frequently asked question is that of expected keyboading speeds. This is 
one of the reasons for the creation of the National Composition Association 
Production Measurement Committee which is finalizing its report on 
keyboard production standards. It will be interesting to compare some of the 
conclusions of the Production Measurement activities against your own 
findings. While these standards are not based on a controlled study and do 
not indicate the specific type of equipment used, they do offer some 
interesting information. 

It is obvious that a manual linecaster offers some mechanical limitations 
which affect the productivity of the equipment. The Monotype and 
TeleTypesetter keyboards also offer some mechanical limitations along with 
end-of-line decisions which can hamper total productivity. In these 
standards, the only apparent difference between typewriter keyboard (non- 
justified tape) and straight typewriter is the typing skill of the operator. This 
confirms some of the preliminary conclusions drawn from the NCA study. 
In fact, we have suggested that all operators be given a typing test in order to 
ascertain their production potential. A copy of the typing test is available 
from NCA/PIA headquarters. 


84 Composition Input 



7. Optical character recognition (OCR) systems 


Before the typesetting and printing process can begin, manuscript material 
must be converted. This conversion process embraces two very important 
functions: 

-it converts the material to a form compatible to the typesetter 
being used. In practice, this usually means the creation of a 
punched paper tape in the appropriate 6 level TTS format. 

-it adds the typesetter control instructions, so that the finished 
textual material is of the desired format, and appears in the 
right font and size. 

There are conversion processes that accomplish both steps by a single 
conversion operation, while other processes require two separate steps to be 
taken. 

The actual choice of the conversion process will depend to a great degree on 
the form of the manuscript, or document. Where the manuscript is in the 
form of a typed page, it may be possible to read by pages directly by using an 
OCR system. Or, the pages may be converted to tape by using keyboard 
devices. Where the manuscript is in the form of machine readable tape, it 
may be possible to use this tape to create useful typesetter input. 

Extracting information from a document falls under the general heading of 
character recognition. This idea was first developed by the data processing 
world in an attempt to eliminate the input conversion step. It was found that 
many documents were generated by a machine to impart information to a 
human, and that this same information was subsequently required by a 
machine. A primary example of this problem is the utility bill or credit card 
bill. A slightly different problem exists where it is necessary to give unique 
identity to a document, such as a check. In all these cases, the general 
principles of character recognition were promoted so that it was not 
necessary to use a keypunch operation to get this material back into machine 
readable form. As the volume of such keypunching increased, the value of 
character recognition equipment also increased. 

A number of machines are now available to address certain specific tasks, 
such as reading journal tapes from adding machines, reading sales slips 


Composition Input 85 



embossed by credit cards, reading hand marked test scores, and so on. A 
number of these machines have important constraints on the number of 
characters they can recognize, the type of material upon which the 
characters are printed, the size of the document they can accept, etc. 

In general, two basic recognition schemes are being used today (1) reading 

the character by magnetic means, or (2) reading the character by optical 
means. 

The most widely used magnetic method of character recognition is called 
MICR (Magnetic Ink Character Recognition). In this scheme, the 
characters are printed using special magnetic ink. The characters, due to 
their distinctive shape, have unique magnetic properties, and it is possible to 
read these characters by using a special magnetic reader. The character set is 
limited, consisting only of the numbers 0-9 plus 4 special characters. 

The banks are the largest users of this concept. Printed checks usually 
contain the name of the account holder (be it an individual or a company) 
and the account number using MICR characters. In this way, the check can 
be machine read to establish the account number. 

The associated code is a machine readable code associated with a human 
readable character. In this way, the information is conveyed to the human 
reader by conventional characters, while the machine reads the associated 
code. A great deal could be said about the applications of the different 
techniques, but our primary purpose is to simply expose the concept. 


Datatype uses an IBM Selectric typewriter equipped with a special ball. It is 
also possible to modify a line printer to produce this font, but the primary 
value of the idea is the ability to use a standard office typewriter as a data 
recorder. Once the page is produced using the Datatype code, it can be read 
by a special page reader made by Datatype. The reader scans the page, a line 
at a time, and converts the information to a punched tape or magnetic tape. 
The machine readable code located under the human readable character is 
similar in principle to the Semagraph, and shares the drawback of being 
limited in the number of codes available. Datatype can print 88 codes, and 
since some of these codes are related to control, they do not print a human 
readable character. This has the effect of reducing the character set of the 
typewriter. The Datatype page reader is in the $10,000 class, and the reader 
is used in conjunction with a minicomputer and tape punch or mag tape drive 
to form the entire system. 


The Potter approach uses bits printed both above and below the human 
readable characters. Potter has based the system on the use of cards, which 
can be printed by means of a special typewriter (either Selectric or type bar) 
or a special line printer. The value of the system seems to be in the ability to 
sort and collate cards. The card reader is available with interfaces to permit 
connection directly to a computer, thus the need to convert to a tape is 
eliminated. The reader is rather expensive, being in the $20,000 range. The 
character set is limited to 64 codes. 


86 Composition Input 



Dual Image uses a full 8 bit code pattern printed below the human readable 
character. The images are generated by a special printer, which can print a 
character set of 128 different symbols. In addition, the 8 bit code pattern can 
furnish any of the 256 different possible codes. The printer can also produce 
two character mnemonics using a single character space. The printer 
produces a strip of tape, which is then read by a very simple reader in the 
$1,000 price range. Interfaces are available to connect directly to computers, 
so there is no need to convert to a different machine readable media. The use 
of 8 bits in the code pattern also permits the printed code to be in machine ' 
compatible format, so that no code conversion is required. 


Mark Sense requires the use of a special mark placed in a particular zone on 
the document. The mark is usually made by a soft lead pencil, and the reader 
merely determines the absence of or presence of this mark. The concept is 
widely used in scoring tests, where the answer may be “true or false” or 
possibly one choice in five. The business world uses a similar scheme, but it is 
based on a numeric only technique where the mark is placed in a one out of 
10 basis. Mark sense readers are marketed by IBM, NCR, and Hewlett- 
Packard (to name a few) and start as low as $3000. 


OCR, itself, recognizes the character by its shape, similar to the way the 
human eye reads. OCR readers are classified by their ability to read 
different font styles or character shapes (single or multifont), and by the 
number of different characters that can be recognized (sometimes referred to 
as the “vocabulary” of the reader). 


Single font readers are manufactured to read a specific font style, such as the 
USASI OCR-A font, or the European ISO OCR-B font. 

Since these fonts are stylized, they must be generated by typewriters or line 
printers that are so equipped. In general, the spacing between characters, the 
spacing between lines, and the reference to a given side of the page must be 
rather precise. OCR readers are frequently sensitive to paper material, since 
the “reflectivity” of the paper vs the reflectivity of the image must be 
consistent. 


Multi font readers have the ability to be programmed to read any number of 
different font styles. They usually have a fairly large vocabulary, and can 
read upper and lower case characters, the number set, punctuation marks, 
plus some special characters. Some machines can even read carefully 
executed handwriting. In the main, these machines can accept material that 
has been produced on standard office typewriters or line printers. Obviously 
the more capability of the machine the higher its cost, and large vocabulary- 
multi font readers are very expensive. 


As with single font OCR readers, the multi font system reads the input page 
and produces a machine readable tape, either magnetic or paper. Since a 
computer is included as an integral part of the system, it is possible to do 
some formatting of the output material. The nature of the formatting, and 
its extent, is a function of the computer program. 


Composition Input 87 



Document readers can generally read one-to-five lines of information from a 
paper coupon, stub card, or similar document. Most document readers can 
handle documents ranging in size from 2x4 inches to about 4x8 inches. 

Document readers are widely used in reading turn-around forms such as 
statement remittance stubs where the printed output from a computer later 
becomes input to the computer. 

Page readers are generally designed to read large and variable amounts of 
alphanumeric information typed or printed in normal page format. Most 
page readers accept sheets from 8 b x 11 to 12 x 14 inches in size. Page 
readers also have the capability of reading somewhat smaller sheets and 
continuous fan-fold or rolled sheets printed by computers or by specially 
equipped typewriters. 

Bar-code readers sense marks that are used in combinatorial form to 
indicate data. The type of marks used varies, but in most cases the marks 
cannot be formed by hand and are not easily readable by humans. 

Usually, special devices are required to produce the bar code imprinting. Bar 
codes also suffer because the code and the character occupy a good deal of 
space. 

Character readers are the upper-class of optical readers. They translate 
human-readable characters into machine-readable form. Many specialized 
fonts have been developed to simplify the character recognition logic and 
hence lower the price. 

A document reader reads documents of less than standard letter size (8.5 x 
11 inches). A page reader reads at least letter-size documents and usually 
larger ones. Another way of distinguishing between document and page 
readers is that document readers generally read one or two lines per 
document, while page readers can read many lines from each document. 
Single font means that the reader can be equipped to read one typeface only. 
Multiple font means that the reader can be equipped to read several 
typefaces but only one at a time; switching between type faces can be a 
manual or programmed feature. Multi-font means that the reader can read 
multiple typefaces intermixed; this is the most sophisticated and expensive 
type of optical reader. Journal tape is the rolls of tape used by adding 
machines and cash registers. 

Now that we have defined the types of devices we re talking about, let’s 
discuss their application. 

Who Uses Optical Readers? 

Mark readers are used principally for data collection and for entry of limited 
amounts of data on previously punched cards. 

Bar-code readers and character readers have many applications. The 
principal ones at present are the reading of slips imprinted with a credit card, 
processing of turnaround documents, and sorting of the U.S. Mail. 


88 Composition Input 



Various manufacturers estimate that any installation having anywhere from 
7 to 12 or more keyboards can profitably make use of a character reader. 
Where do the savings come from to pay for this expensive beast? 

One place is the lessened cost of labor. Since manual input for the character 
readers is typically prepared on a typewriter, the hourly wage rate is 
generally lower than for keyboard operators, while the output is higher and 
the rate of errors is lower. The ease with which errors can be corrected when 
preparing typewritten documents contributes to the speed in comparison 
with keypunching. One' user estimates that about 10 percent of the 
documents processed at his installation contain errors detected and 
corrected by the operator. Some readers contain special facilities for 
recognizing a character skip symbol or strike-throughs to further ease 
correction of errors detected by the typist. 

Most optical character recognition systems consist of four basic units. 

Document Transport Unit 
Reading Unit 
Recognition Unit 
Control Unit 


The transport unit moves documents past a scanner that converts the 
characters on the document into electrical signals that are then analyzed and 
recognized by the recognition unit. The recognition unit matches patterns or 
representations received from the scanner against stored reference patterns. 

The transport moves the documents from an input hopper or feed roll past 
one or more scanning units to one or more output stackers. In certain 
equipment the documents are read while still moving, but in most cases, the 
document is stopped and read. Document transports employ combinations 
of vacuum, air blast, and friction to separate and feed individual documents, 
while belts and rollers are used to transport the documents past the scanning 
unit. 

The speed of most OCR systems is limited by the speed of the document 
transport. The scanning unit determines the speed of the OCR unit when the 
amount of data per document is large. 

Functional features or characteristics that effect the complexity of the 
document transport are: 

Detection of double documents and jams 
Detection and correction of document skew 

The scanning unit converts the printed information on the document into 
electrical signals that will enable the recognition unit to recognize the printed 
characters. The five distinct methods currently in use for converting optical 
signals into electrical signals are: 

1. The rotating disc scanner uses a high quality lens system to project light 
reflected off the document onto a rapidly rotating disc. The rotating disc has 


Composition Input 89 



apertures extending from the center of the disc to its periphery. Behind the 
character image area on the disc is a fixed plate containing a single aperture. 
This aperture is so oriented that each aperture on the rotating disc 
successively intersects along the entire length of the fixed aperture as the disc 
rotates. 

The rotating disc scanner reads one character at a time. Movement from one 
character to the next character or from line to line is accomplished by 
repositioning the lens system or by moving the document. Therefore, this 
type of scanner is relatively slow in comparison to other scanning methods 
mentioned. 

The advantage of the rotating disc approach is that it is relatively simple, 
permits paper to be exposed to ordinary light (actually the more light the 
better), requires only one or, at most, a few photocells, and permits 
adjustment for different background colors by varying the threshold voltage. 


The disadvantages of this method are that high-speed discs are noisy and 
difficult to manufacture and the throughput rate of the system is limited by 
the disc revoltion speed. It appears that about 400 characters per second is 
the upper limit for OCR systems employing this approach. Therefore, the 
rotating disc is being replaced by faster methods of scanning for a number of 
applications. 



90 Composition Input 




2. The flying spot scanning method uses a cathode ray tube (CRT) to 
generate a small (spot) of light that is projected onto the document being 
read via a lens system. The document must be located in a lightproof 
compartment where the reflected light can be picked up by one or more 
photomultipliers. 

The CRT light beam is swept across the character in a raster-type screen by 
the CRT control logic. The beam can be moved very rapidly by the control 
unit to any location on the document. This ability enables a flying spot 
scanner to locate a line of print anywhere on the document without having to 
read the entire document. This positioning capability may also be used to 
follow the lines of a character, thus making it particularly adept at reading 
multiple font and hand printing. 

3. In a photocell scanning system, a high intensity light source illiminates the 
document which is in motion and the reflected image is focused onto a 
grouping of photocells or light pipes which feed the photocells. The grouping 
of photocells can either be vertical, relative to the character being read, or it 
can be a parallel array of photocells. 



In the vertical grouping of photocells, each character is sampled as it moves 
from left to right. The use of a vertical grouping of photocells speeds up 
scanning operations by simultaneously sampling a number of points which, 
when combined, add up to a complete vertical slice of the character. The 
electrical signals generated by each of the photocells are then coverted into a 
binary mode and each slice is stored in shift registers until the entire 
character is sampled. 


Composition Input 91 




The parallel- array approach looks at all points of a character 
simultaneously. The speeds of each method are comparable; however, in the 
parallel or full photocell array approach it is possible to measure analog 
information completely. This capability enables different shades of black 
and white to be read and thus provides a greater probability of recognizing a 
smudged or dirtied character. 

4. The vidicon technique projects the characters to be scanned onto a vidicon 
television camera tube. The vidicon tube is instantaneously exposed to the 
characters (in camera fashion) by either flashing a light (flash tube) on the 
document when the characters are ready to be read, or the document is 
constantly illuminated with a strong light and a high speed electro¬ 
mechanical shutter is used to “snap the picture.” The image on the face of 
the tube is then scanned by an electron beam which generates an electrical 
analog signal. The resulting signals are quantitized to digitally indicate black 
or white. 

The vidicon scanner can store a group of about 45 characters on the face of 
the tube and, therefore, documents containing this number or fewer 
characters do not have to be moved during the scanning operation. With the 
development of much higher resolution vidicon tubes, it would be possible to 
store the entire document and eliminate mechanical movement completely 
during the scanning operation. Vidicon scanners are presently classified as 
medium speed (500) characters per second. 

After the scanning has taken place, an electrical representation of the 
character is transmitted to the recognition unit which then identifies the 
character. With some systems there is an intermediate prerecognition step in 
which undesirable electronic “noise” caused by white spots on black ink, 
dirt, or inadvertent ink spots is reduced. The value of this technique is still in 
dispute. The most common types of recognition units currently in use are: 

1. The use of optical masks is one of the earliest recognition techniques. It is 
based on the use of one or more photographic masks for each character. An 
attempt is made to measure how well the character projected matches with 
the mask. 



92 Composition Input 



Photocells behind the mask measure the total light passing through tlje 
mask. Ideally, no light should pass through the mask if it matches the 
character being identified. In practice the match is usually not precise 
enough to blank out all the light so a threshold value is established as a 
tolerance. 

This technique has the ability to identify a full alpha-numeric character set, 
however, small differences in character shape may cause character 
identification errors and high reject rates. No known commercial OCR 
systems are currently using this method. The concept has the potential for 
providing low-cost OCR systems if the input data can be closely controlled. 

2. Matrix matching is a widely used recognition technique which stores 
electrical signals received from the scanner in a digital register that is 
connected to a series of resistor matrices. Each matrix represents a single 
reference character. Each resistor matrix is connected to a second digital 
register which contains a voltage representation of the character. The 
voltage of the scanned character is compared with the second digital register 
and the resistor matrix. 


Recognition is based on the comparison of the voltage representations in the 
two-shift registers. This recognition technique is well developed and can 
handle a complete alpha-numeric character set and is easily modified to 
identify characters from several type fonts. 

3. Analog waveform matching is a recognition method that has been used for 
some time, particularly in the magnetic character reader used by the banking 
industry. This method is based on the principle that each of certain 
characters passing under a read head will produce a unique voltage 
waveform as a function of time. Characters are identified by matching their 
waveforms against reference waveforms. 

The major disadvantage of this technique is that only a small number of 
characters have easily identifiable waveforms, thus limiting this application 
to the reading of only numerics plus a few special symbols. Machines using 
this technique have reading speeds of approximately 500 characters per 
second. 

4. Frequency analysis is a digital recognition technique developed for fonts 
using vertical lines. The CMC-7 font printed ith magnetic ink, is the most 
widely used example of this technique. The width of the gaps between the 
vertical lines of a character form a code that is unique to the character. 
Characters can be identified by comparing the frequency and number of the 
wide and narrow gaps with the stored codes for each alpha-numeric 
character. The advantages of this technique include the ability to handle a 
full character set. 

5. Stroke analysis or feature analysis is a recognition technique based upon 
the differentiation of characters by the number and position of vertical and 
horizontal strokes or lines. The formation of the unknown character is 
matched by a special purpose computer against stored truth tables 
representing each reference character. The capability of this technique has 


Composition Input 93 



been increased to the point where hand-printed numerics, and several fonts 
can now be recognized. 

The systems control unit performs data editing and formating, identifies and 
interprets various formats for different documents, sequences some systems 
operations, and provides the interface required to record data on an output 
device such as magnetic tape or punched card. The systems control unit may 
be a special or general purpose computer, or a plugboard. 

In most of the early optical characters readers, systems control functions 
were performed with plugboards that greatly limited the system’s flexibility 
and data processing capabilities. While some readers still use a plugboard, 
particularly mark sense and document readers, most current OCR systems 
and those expected in the future use a combination of systems software 
capability and a general purpose computer for control functions. The 
computer can be a mini-computer supplied as an integral part of the system 
or the reader can be operated on-line to a larger data processing system. 

OCR devices with computer logic capability and specialized software are 
performing many sophisticated data processing functions such as validation 
of self-checking numbers, reconstruction of missing digits, and identification 
of characters by context analysis. 

There is not too much to say about recognition of marks by a mark reader. 
Typically, they are diagonal slashes made in a preprinted box or outline. 
Care must be taken when erasing because if the paper is roughened too 
much, it will have a low reflectance and make the reader conclude that the 
roughened area is a mark. In general, more care must be taken with erasures 
than with the older conductive mark-sense technique. 

The style of the type face printed is called a font. The group of symbols that 
the reader will recognize is referred to as the character set. Note that a 
particular reader may not read all the symbols of a particular font. The usual 
situation is that only the numerics are recognized of a font that also contains 
alphabetic letters. Occasionally, a larger set of characters can be recognized 
than are in the font proper. 

There are many fonts that are used today in addition to traditional printing 
type styles. A few of the more sophisticated readers recognize printing type 
styles and in addition to or in place of the OVR printing type styles in 
addition to or in place of the OCR fonts. 

The more commonly used OCR fonts include: 

• NCR NOF (Numeric Optical Font) — This is a numeric font, usually 
imprinted by an adding machine or cash register. It is widely used in retail 
applications. 

• ANS I and IV — Previously called OCR A and C, these fonts were 
developed by the American National Standards Institute and are likely to 
become the most widely used OCR fonts. The two represent two sizes of the 
same typeface. For perspective, the sizes are roughly equivalent to 10 point 
(I or A) and 14 point (IV or C) type. Imprinting devices abound, including 


94 Composition Input 



the IBM Selectric Typewriter, most other major brands of electric 
typewriters, and some line printers. 

• OCR A and C — See ANS I and IV, above. 

• ISO B Popular in Europe, this font is now under consideration for 
standardization in the United States. It differs considerably from the ANS 
standard and is characterized by being closer to conventional type faces than 
the ANS fonts. Exponents of the ISO B font are concerned about readability 

by people, even though it is more difficult to build the machine recognition 
logic to handle it. 

• Farrington 7B, 12F, and 12L — Another popular group of OCR fonts, due 
to Farrington’s early appearance on the OVR scene. The three codes have 
somewhat similar shapes but differ in size and character set. The 7B and 12F 
are numeric fonts, while the 12F is alphabetic only. The 7B is much larger 

than the 12 F/F. Imprinting is normally done by a special typewriter or a 
credit card embosser. 

• IBM 1428 This is an alphanumeric font associated with the IBM 1428 

Optical Reader and imprinted by an IBM 1402 Fine Printer or IBM 
Selectric Typewriter. 

• IBM 407 — A font produced by the widely used IBM 407 Accounting 
machine. 

• E-13B — This is not properly an OCR font. It was developed by and for 

banks prior to the development of OCR. It is a highly stylized numeric font 

intended for printing an magnetic ink to facilitate the sorting and processing 

of bank checks. It has not caught on anywhere else, but most banks use it. 

Some optical readers can read this font, which enhances their suitability for 
banks converting to OCR. 

All fonts, both numeric and alphanumeric, usually include a few special 

symbols for control purposes. The OCR A and C fonts include a full array of 
punctuation symbols as well. 

The scanning technique is the method for optically converting the printed 
images to electrical signals. Some sort of photosensitive device, photocell or 
phototransistor, is used to sense the light reflected from the document. For 
bar-code and character readers, additional components are required to scan 
portions of the code or character in proper order so that the features, and 
thus the character, can be identified. The scanning components can be an 
array of photo devices, a mechanical disc, or a flying spot (CRT). The 
photo-device array is used by most bar-code readers. Size normally 
interferes with using it for scanning characters, but note that REI does quite 
nicely with it. The mechanical disc technique employs a rotating disc with a 
slit in it to project a beam of light over the character in a predetermined 
order. The Hying spot scanner uses an electron beam that is moved within a 
CRT to generate a spot of light, thus providing, potentially anyway, a much 
faster scan rate. The Hying spot scanner is very adaptable to reading multiple 
lines, while the mechanical disc scanning technique requires either an 

incremental document transport or an elaborate system of mirrors to scan 
multiple lines. 


Composition Input 95 



Once the printed image has been translated into electrical signals, the 
recognition logic interprets these signals as a particular character. 

Three principal recognition techniques are used for character interpretation: 
matrix matching, stroke analysis, and curve tracing. Matrix matching 
involves comparing the matrix of signals caused by the reflecting (paper) and 
non-reflecting (printed character) portions of the character with a set of 
signals for each character until a match is found. Typically, the closest 
match within prescribed limits is identified, because a perfect set of signals is 
most unusual due to variations in print and paper quality. (The stroke 
analysis method is somewhat similar on a much simpler basis.) Readers 
employing this technique usually are reading a highly stylized font 
specifically designed for the technique. Curve tracing logic actually traces 
out the outline of the character to derive a set of signals for analysis. The 
curve tracing technique is adaptable to variations in character size and 
orientation, making it a good choice for interpreting hand-printed 
characters. However, breaks in the printed character tend to affect this 
method more than the matrix matching method. 

The purpose of the optical reader is to generate data in a typesetter - 
readable form. Magnetic tape, punched cards, and punched tape are 
conventional computer-readable forms. De facto standards, now official, 
have been established by IBM for magnetic tape and punched cards! 
Teletype has done essentially the same for punched tape. Exceptions that are 
relatively new on the market are magnetic tape cassettes and 96-column 
cards — but neither of these have had much impact on the optical reader 
market, for the simple reason that they haven’t had much impact on 
computers as yet. 

Readers that handle one size of documents are easy to rate for performance 
because it is predictable. In a similar manner, a journal tape reader typically 
transports the tape at a fixed rate, with predictable performance. Readers 
that handle different sizes of documents and variable-size data fields are not 
as conducive to having their performance stated in simple terms. 

Three ways of measuring the performance of optical readers are documents 
per minute, lines per minute, and characters per second. The documents-per- 
minute rating is usually most applicable to mark and bar-code readers, as 
well as to character readers that read only one or two lines. The lines-per- 
minute rating is usually most meaningful for journal tape readers. The 
instantaneous character scanning rate in characters per second is probably 

the most meaningful single measure for character readers that read whole 
pages of text. 

Careful evaluation of timing information, which often becomes quite 
complex, is necessary to accurately predict the performance of the more 
sophisticated character readers. The size of the document, the amount and 
location of data on the document, and processing of the data read can all 
affect the rate at which documents proceed through the reader. One-line 
units can be affected by other activities of the computer, if running in a 
multiprogramming environment, or by poor programming of input/output 
functions. 

Reject rates of 0.25 to 0.5 percent cause some users to become startled and 


96 Composition Input 



disillusioned when they see 30 or 50 percent of the documents going into the 
reject pocket. 

There are three principal types of errors of concern to users of optical 
character readers: ambiguous characters, invalid data, and documents in 
poor condition. 

Ambiguous characters are those for which the reader cannot make a 
decision about what character each should be. There can be many reasons. 
Typical ones include broken or poorly formed characters and dirt or other 
marks that are picked up by the reader. Handling of this situation varies 
with the reader and with programming. Many readers automatically rescan 
an ambiguous character. Some substitute a standard character for all 
unreadable characters and continue. Others display the character on a CRT 
screen for operator determination; sometimes adjacent data is also displayed 
to give the operator more context for making the decision. Printing quality 
and paper quality can drastically affect the incidence of this type of error. 

OCR users quickly learned that the inclusion of checks in the data was 
extremely useful for insuring the maintenance of an adequate throughput 
level by reducing the number of rejected documents. This technique most 
frequently takes the form of repeated data fields, particularly for numeric 
entries. The technique is applicable only if the data can be processed and the 
actions of the reader controlled on the basis of the result. 

Another commonly employed check is the check digit. The digits of a 
numeric field are manipulated, and there are several standard formulas, to 
generate a check digit. This digit is included in the input. The reader or 
associated processor generates another check digit while reading and 
compares it to the one read in. Failure of this check normally causes the 
document to be rejected. 

Documents that have extraneous items on them (such as stamps) or that 
have been badly mutilated can cause misfeeding and/or jams. The typical 
character reader is far less susceptible to this kind of jam than the average 
card reader, but people have been conditioned not to fold, spindle, or 
mutilate punched cards. 


To permit processing of the manuscript by an OCR system, the typed pages 
must satisfy the criteria of the OCR reader. Some of these constraints are: 

-the type style must be one the OCR reader "knows” 

-the character set should not exceed the character library 

-the paper or document material must be acceptable to the reader 

-the line length, and line placement, must be appropriate for the reader 

The first step is to read the character, and this is done by using scanning 
techniques. A single point of light is moved over the character image, and the 
light and dark portions of the scanning pattern are fixed on a matrix. The 
dark portions of the image represent dots on the matrix, and it is then 
possible to assign X and Y co-ordinate points to the dots. Thus, the 
character image is converted to a dot pattern, and each of the dots can be 
represented electronically as a result of the X-Y co-ordinates. 


Composition Input 97 



OCR readers scan several hundred lines in each of the two planes to achieve 
fairly high levels of resolution. The process of reducing the image shape to a 
dot pattern on a matrix is sometimes called "digitizing”, or converting the 
image to digital representation. 


Scanners are either mechanical or electronic, with the mechanical scanners 
controlling the movement of the light point by means of mirrors. Some 
mechanical scanners position the document on a drum, and rotate the 
document to derive one plane, and a mirror establishes the other plane. 
Electronic scanners use CRT principles and control the light point by 
deflection plates. 

Once the character image has been digitized, it is then passed on to the next 
stage, where the character being read is compared to a library of all known 
image patterns. This is called correlation. When a reasonable comparison is 
achieved, the character is identified. 

The identified character is now passed on to the format and control element, 
where the material is organized for proper recording on the output tape. The 
control and format section also takes care of such things as discarding 
characters that had been deleted. Some systems also utilize a small TV 
screen, so that if the reader scans a character, but can not find a comparison 
in the character library, the unknown character is displayed for the operator 
to see. The operator can then interpret this image, insert the correct 
character back into the system via a keyboard, or take other appropriate 
action. 

The output of OCR systems is either magnetic tape or punched paper tape. 
The higher speed machines use magnetic tape because of the higher 

recording speeds, although paper tape can also be used if the speed sacrifice 
is not important. 


OCR has developed very slowly as evidenced by the fact there are currently 
only about 1,500 OCR installations versus about 80,000 computer 
installations, even though the first commercial OCR system was installed in 
January 1956, only three years after IBM shipped its first commercial 
computer. The primary reasons for slow growth of OCR were the lack of 
flexibility of the first generation OCR systems and limited view of the 
potential OCR market by the industry and potential users. Until recently, 
almost all OCR devices were limited to reading short vertical or horizontal 
dashes (mark sense), or highly stylized fonts designed primarily for OCR 
reading. The lack of flexibility of earlier OCR systems has been remedied 
and most OCR systems introduced since 1968 and those scheduled to be 
introduced in the near future are capable of reading a number of different 
type fonts, including hand-printed numerics plus a few symbols. In addition, 
the software programming has been greatly improved giving the systems 
considerable editing and formatting flexibility. 


Optical character recognition systems can be classified by the complexity of 
the font selection which the machine can handle and by the physical 
characteristics of the media presented for reading. 


98 Composition Input 



OCR systems fit into the following categories when classified by font 
selection: 

1. Optical Mark Readers 

2. Stylized Font Readers 

3. Multifont Readers 

Optical mark readers are an improvement in the mark sensing techniques 
where the location of graphite pencil marks was determined by measuring 
the electrical conductivity of the pencil mark. Now, most mark reading is 
done optically. At one time, optical mark readers were used primarily in 
scoring tests and questionnaires and in survey applications. Today, they are 
also used as data acquisition devices in payroll, inventory control, meter 
reading, and a number of similar applications. 

The scanning systems of optical mark readers and optical character readers 
are basically the same, but their recognition systems differ greatly. Both 
systems use a set of photocells to detect a drop in reflected light caused by a 
mark or character on the paper. The recognition system for an optical 
character reader is considerably more complex than that of an optical mark 
reader. An optical mark reader recognizes only the drop in light and 
determines the value of the mark by noting its position in relation to some 
reference point such as the beginning of a line. An optical character reader 
notes not only the gross drop in light but the coordinates on a two 
dimensional grid at which the drop occurred. 

There are approximately 50 major type fonts used by computer printers and 
typewriters. This large number of different fonts has presented a great 
restraint on the development of optical character recognition systems. In 
order to control data input, the initial manufacturers of OCR equipment 
created special stylized fonts that were compatible with the design of their 
own equipment. This lack of standardization during the past decade has 
severely limited the growth and acceptance of OCR as an input technique. 

A number of type fonts have been developed over the past several years for 

machine reading. The USA Standards Institute has recommended a type 

ont that is generally suitable for typewriters and is also suitable for OCR. 

The USASI OCR-A font is the most common font, and is read by almost all 
OCR systems. 


Composition Input 99 



The following pages are excerpted from the Olivetti 
manual on Optical Character Recognition fonts and 
keyboard arrangements. 



This booklet illustrates the most common Editor 2 key¬ 
board arrangements for use in Optical Character Re¬ 
cognition applications. 

When ordering an OCR Editor 2*» the following questions 
must be answered: 

1. Uhat character set is the Optical Reader pro¬ 
grammed to recognize? 

2. Uhat Keyboard will the application require? 

3. Uhat Keyboard modifications will be required 
due to special symbols such as the character 
delete? 

M. Uhat size carriage will be required to accommodate 
the form size? 

5. Will the forms be continuous or cut sheets? 

b.* Uill carbon or fabric ribbon impressions be re¬ 
quired? 

Use the following information to obtain answers to the 
above questions. 

Character Set 

A keyboard number is located in the space bar on each key¬ 
board diagram. This numberi used when ordering** describes 
both the keyboard arrangement and the character set. For 
example-* keyboard 2-^03-! indicates a 10-Key Keypunch 
cluster keyboard arrangement and the USA Standard Character 
Set • 

A type sample precedes the diagrams of the keyboards avail¬ 
able with each character set. The Optical Reader will de¬ 
termine the character set that is required. 

Keyboard 

After determining the character set** the keyboard should 
be selected- 

The keyboard should be chosen to meet the needs of each 
application. A description of each keyboard has been in¬ 
cluded for this purpose. 


V 



100 Composition Input 



Keyboard Modifications 

The special symbols will also be determined by the Optical 
Reader. The most frequently used character combinations 
are listed on page 3 • Others are available upon request. 

Any keyboard may be modified to include special symbols*! 
when required. 

Carriage Length 

The size of the form will indicate the length of the 
carriage to order. 

Pin Feed Platen 

When continuous forms are used*, a pin feed platen may 

be installed on the Editor 3. Refer to the current 

Technical Information Bulletin I TIB > for the ordering 

procedure*, prices and installation data. Additional 

information is contained in the Features and Attachments 
Booklet. 



3 


The maximum form widths 

Carriage length 

13" 

17" 


for each pin feed platen are: 

Maximum form width 

11 7/A" 

IS 7/6" 


Ribbon 


Most Optical Readers recognize 
carbon ribbons. 


impressions produced by 



Composition Input 101 


USASI * OCR type style and keyboards 


THIS IS A SAMPLE OF THE UNITED STATES OF AMERICA 
STANDARDS INSTITUTE-, USASI*-, OPTICAL SCANNING TYPE FACE. 

ABCDEFGHIJKLMNOPfiRSTUVUXYZ 
15345^76^0 YJMSX | »*{>_- + = — 

Optional character combinations 

j. 2 b a * j i a: i 

others available upon request. 


♦Previously known as the American Standards Association (ASA). Identified internationally as OCR-A. 


Upper case alphabetic USASI characters are located in both shift positions on keyboard 2-900. 
Less shifting and greater typing speeds are the advantages of this keyboard arrangement. 



\ EDITOR 2 

USA STANDARD CHARACTER SET 

OCR #2-900 j 

d ... 3 


Keyboard 2-905 is used with optical readers that recognize different symbols from those on key¬ 
board 2-900. Both are used for the same applications. 



I EDITOR 2 

USA STANDARD CHARACTER SET 

OCR #2-905| 

r -j 


102 Composition Input 















































Keyboard 2-903-1 is used when the OCR typing application contains a large amount of numeri¬ 
cal information. It contains upper case alphabetic USASI characters and the 10-key keypunch 
cluster. Numerics can be entered quickly and easily with this keyboard arrangement. 


IBII131IIBIIHUBIIH 


IIBIIIBIIBIII 


EDITOR 2 USA STANDARD CHARACTER SET 


OCR #2 903-1 


Keyboard 2-907 is used when typing a large amount of numerics on forms that do not have 
pre-printed field separators. This keyboard differs from keyboard 2-903-1 by containing field 
separators in both shift positions in place of the group delete. The group delete replaces the 
underscore. Less shifting and greater typing speeds are the advantages of this keyboard 
arrangement. 


MAAG4N 1 

1 MAAGIM 

&YTASS I 

1 « T 


IBIIBIII3IIIBIIIBIIIBI 


EDITOR 2 


USA STANDARD CHARACTER SET 


OCR #2 907 


Keyboard 2-906 is used with optical readers that recognize different symbols from those on key¬ 
board 2-903-1. Both are used for the same applications. 


TA® I I TA® 

CUAft II SCT 


EDfTOR 2 


USA STANDARD CHARACTER SET 


OCR #2 906 


Composition Input 103 










































































Keyboard 2-956 contains upper case alphabetic USASI characters and the 10-key keypunch clus¬ 
ter for rapid entry of numerics. It is recommended only for accounts using competitive typewrit¬ 
ers to maintain keyboard uniformity. 



Keyboard 2-965 contains upper case alphabetic USASI characters and the 10-key keypunch clus 
ter for rapid entry of numerics. It is recommended only for accounts using competitive typewrit 
ers to maintain keyboard uniformity. 


H| bi | ea f mmmmmmmm 

1 i TA * 1 V TA# 1 

| | CXJLA* || S€T I 



iisiliiilisiiisiliiiliailu 

!■ IBIIDI |g|||g|[|g|||g| 
tel lOlllOIIIBIIIBlIlBI Id 

1ISII■ IlSlIlglllHIES 

b o | a mmmrn 

pa ii hiibhibiiimJ 

ASJ151151 ISillSlIli 1 

gj[B B M ||tJj|[Og 

| EDITOR 2 USA STANDARD CHARACTER SET OCR #2 965j 


104 Composition Input 






























































This type style is used for general correspondence 
by combining the lower casei illustrated in this 
samplei with the USASI type face- 

ABCDEFGHIJKLflNOPflRSTUVUXYZ abcdefghijklmnopqrstuvwxyz 
15345b7ATD YiPrl** | &* O ?/* 


Keyboard 2-901 is used in two ways. The upper case alphabetic USASI characters are used for 
OCR typing. For general correspondence typing, use this keyboard in the conventional way. 




Composition Input 105 










































































106 Composition Input 












































8. What a computer does 


The Fall of 1962 saw computers utilized in typesetting for the first time. 
Both the IBM 1620 and RCA 301 were general-purpose computers 
programmed to accept non-justified paper tape input and produce justified 
and hyphenated paper tape to activate linecasting units. The perforator 
operator typed an “endless” or “idiot” tape with no end-of-line decisions; 
however, codes for font, measure, quadding and other typographic functions 
were encoded. Essentially the computer performed electronically what the 
counting perforator and linecasting spaceband did mechanically. The 
computer added the important function of word division. Two methods were 
used: logic (or probability) and dictionary (or lookup). Here is an example of 
a probability program based on certain rules of logic. This limited but 
somewhat effective program is used in Compugraphic phototypesetting 
units. 

Photo Unit Hyphenation Program 

With the hyphen switch in the ON position the following sequence must 
occur before a line ending decision is made where the machine will 
automatically insert a hyphen: 

1. The machine must have read at least two letters in a word. 

2. A letter in the Kb group must be followed by a letter in the Kc group: 

Kb - B, C, D, F, M, N, P, R, T, V, X, Y 

Kc - B, C, D, F, J, M, N, P, S, T, V, W, Z 

3. That there are at least 2 letters following the hyphen location to bring 
over to the next line. 

4. That the word the hyphen location is found in will meet the requirements 
to make a line ending decision, ie., that there are enough characters and 
bands for justification. 

5. That the insertion of the hyphen does not overset the line with band 
expansion at minimum. 

Discretionary Hyphenation 

Discretionary hyphenation requires the operator to insert a tape feed 
between letters in a word that is a good hyphenation location. The machine 
will insert a hyphen in this position if it is a good line ending point. 


Composition Input 107 



Dictionary programs stored a large number of words in the computer’s 
memory and compared them with words at the line ending point. Correct 
break points were indicated. An “exception dictionary” is a limited list of 
proper names and other words that do not adhere to the rules of logic. The 
computer thus reduces the input burden by eliminating the need for an 
operator’s end-of-line decisions. 

A general-purpose computer may be programmed to perform a variety of 
tasks, such as accounting or billing. A special-purpose computer does only 
one job, such as justification. One of the first such special-purpose 
computers was the Compugraphic Lineasec, which required a monitor to 
make the end-of-line decision for words appearing on a CRT screen. The 
next and much more popular version was the “automatic” Justape (also 
from Compugraphic). Here, the raw tape entered one side and a justified 
tape exited on the other. 

Computer programs may be as simple or as complex as the typesetting 
device and job to be set require. A machine such as the Photon 532 with 32 
typefaces and 23 sizes requires extensive programming to store all the 
character width valves and to “mix” typefaces and sizes in the needed 
format. A format is a repetitive layout, such as a grocery ad or a book page, 
that may be programmed into the computer and accessed via a few 
keystrokes. Characters input after a format command are “automatically” 
set in the desired face, size and position. 


Key to Computerized Composition 

Word division and hyphenation are deceptively simple in theory - simply 
divide between syllables. But it becomes complex in practice because of 
preferential breaks, rule of syllabication and exceptions to these rules, and 
exceptions to the exceptions. The collision between the strict logic of 
computers and the preferences and exceptions of word divisions has 
presented computer programmers with one of their most challenging 
problems. 

The simplest approach was taken by the Compugraphic Linasec. This 
machine relied on a monitor to select hyphenation points when a machine 
was unable to end a line on a word space. According to the Linasec 
reasoning, many typesetters will not be able to recoup the cost of the large 
general purpose computers required to divide words automatically - their 
composition volume will not be big enough. Therefore, typesetters need 
computing equipment that handles only the specialized work of typesetting, 
leaving the work that requires general-purpose capability - hyphenation-to 
the best general-purpose computer of all, which is a man. Then came the 
mini’s! 


Elements other than the word breaks themselves often enter the problem, 
such as a case that one programmer encountered. In this instance he told the 
computer that a widow of three characters was permissible - whereupon the 
computer produced a line containing only a period and two closing quote 
marks. 


108 Composition Input 



This problem, of course, was easily solved by adding extra instructions to the 
program defining what was to be considered a character in this instance. But 
it lengthened the program to be stored in the computer memory, and 
memory capacity is probably the most precious commodity in computers 
used for composition. Additional memory capacity can be added to the 
system, but only at extra cost. 

The choice of program approach often will be dictated by other 
circumstances, as the cases of two newspapers illustrate: 

At a leading Western newspaper, a logic system based on grammatical rules 
of word division was used to guide an RCA 301 computer. It achieves about 
85 percent accuracy in hyphenation - that is 85 percent of the lines ending 
with a word split are correctly broken according to Webster’s first choice. 

Lines incorrectly broken are then reset manually after the type has been cast. 
If the computer hyphenates one line in eight, an accuracy of 85 percent will 
mean resetting 30 lines in 800 (two lines must be reset for every word division 
error), or 3.75 percent of all lines set. 

While corrections of this sort are simple to make in hot metal, they are far 
more difficult in photocomposed film. A Florida newspaper chain had been 
developing film composition techniques for several years, and any computer 
installation they made had to be compatible with film composition. They felt 
85 percent accuracy would create an uneconomical correction problem. 

To increase hyphenation accuracy, they installed an RCA 301 computer 
with auxiliary magnetic tape memory capacity. In this memory they stored 
some 40,000 root words of five characters and up, in groups to speed access 
during operation. Words that break properly according to simple logic rules 
are excluded from the computer’s dictionary. With this system, they can 
achieve a 99 percent accuracy rate with adequate quality for their 
newspapers. This means resetting a maximum of two lines in 800, or a little 
more than .3 percent of the total lines produced. 


In display advertisements and children’s books lines of text rarely end in a 
divided word, since word division would tend to disrupt the reader’s 
concentration on the content. Divided words would be unnecessarily 
confusing to young readers. Some technical manuals and several newspapers 
currently in print also do not contain divided words. In both cases, however, 
the reason for not dividing words is primarily an economic one. If the text is 
prepared manually, the operator has to make certain hyphenation decisions 
which take time and therefore have a cost associated with them. If an 
electronic or mechanical device is used for producing text, these devices are 
usually less expensive if they do not contain hyphenation capabilities. Thus, 
there is definite cost that can be attributed to the division of words, and by 
eliminating hyphenation this cost can be reduced or avoided. 

The alternative to dividing words at the end of a line is to produce lines of 
unequal length by eliminating word division. When these lines are grouped 
together into a column of text, the right-hand margin of the column is ragged 
rather than flush. Such a column is generally considered harder to read 


Composition Input 109 



because the eye must constantly adjust its scanning arc. Therefore, a column 
with a flush right margin should be easier and faster to read. 

The desirability of a flush right margin does not necessarily mean that words 
at the end of a line need be divided. There are techniques in current use for 
producing flush right margins without resorting to hyphenation. One 
technique is to letterspace a word, several words, or a portion of a word in 
the line. The extra space that can be used by a hyphenated word at the end of 
the line is divided up instead and placed between the letters of a word in the 
line. The word or portion of the word is therefore spread apart when it 
appears in print. This sometimes makes the letterspaced word extremely 
difficult to recognize and impairs speed and comprehension. Another 
technique is to take the extra space and insert it in equal amounts between 
the words which comprise the line. This produces larger interword spaces, 
frequently resulting in distracting and unsightly rivers of white space in a 
column of text. Another technique is to divide the extra space on the line by 
the number of characters in the line and to increase the set size of each 
character by this amount. Varying the set size in effect letterspaces each 
word in the line proportionately. Even though the set size of the type may 
differ from line to line, there is so little difference that the distraction to the 
reader is at a minimum. However, this technique is extremely difficult to 
implement. 

Also the larger the column measure, the more interword spaces and 
characters are available into which the remaining space can be placed, 
thereby lessening the need for hyphenation. 

Thus, the primary reason for the prevalence of divided words at the end of a 
line is basically stylistic. Flush right margins are desirable for both ease and 
speed of reading. In addition, the appearance of uniform columns of printed 
matter is very pleasing to the eye. In order to achieve this stylistic treatment, 
the hyphenation of words at the end of some lines is almost always a 
requirement, since the alternatives are graphically unacceptable or incapable 
of being produced on the majority of composition devices. 

There is an additional reason why hyphenation of text is required. Given a 
fixed amount of area into which printed matter can be placed, more text can 
be fitted in if words at the ends of lines are hyphenated. Similarly, a given 
amount of text can be placed in a smaller area if hyphenation is employed. A 
book or newspaper may then require fewer pages, with resultant saving in 
paper or composition. This saving can often outweigh the cost of producing 
hyphenated text, especially since the inception of computerized text 
processing. 

Once the requirement to hyphenate words is imposed, the problem becomes 
one of simulating the human decision process. 

By definition, hyphenation is the division of a word into its component 
syllables. A syllable is defined as one or more letters, constituting either an 
entire word or part of a word, which are pronounced as a single, 
uninterrupted sound. The syllable is the basic unit of pronunciation and 
consists of one single prominent sound and usually one or more less 


110 Composition Input 



prominent sounds. For the majority of English language syllables, the single 
prominent sound is usually a vowel, and the less prominent sounds are 
usually consonants. 

The pronunciation of a word determines the letters comprising each syllable 
in the word. A glance at the pronunciation key in the front of any dictionary 
shows that the same letter can be pronounced in many different ways. These 
pronunciation variations are a problem in any hyphenation routine because 
they are not usually reflected in the spelling of the word, and a computer can 
use only the alphabetic structure in simulating the human decision process. 

A human usually divides the word by pronouncing the word slowly syllable 
by syllable. A computer can simulate this process by scanning the word and 
isolating the vowels in the word. The isolation of the number of vowels in the 
word will not indicate anything about positive hyphen placement; but 
hopefully it will indicate the number of syllables in the word. Since the 
number of hyphen points in the word is one less than the number of syllables, 
this process will yield at least some indication of the number of hyphens to 
place in the word. However, this technique is fraught with problems because 
the same letter in different words may be pronounced differently or not at 
all. For example: 

1. The majority of English words ending in the letter E are pronounced so 
that the E is silent. The words FACE and WRITE both contain terminal 
silent E’s. A computer program that indicates two vowels in each word may 
therefore assume incorrectly that each word contains one hyphen point. The 
problem cannot be solved by always eliminating a terminal E from the vowel 
count because that will produce incorrect results for words like 
ASPOSTROPHE, and MAYBE. The problem becomes more complex 
when the silent E occurs within a word. In the word BASEMENT, the first E 
is not pronounced and therefore should not be considered in a vowel count. 

2. A number of English words contain syllables consisting of several vowels 
and consonants; for example, TONGUE and TORQUE. In both cases, a 
vowel count indicates that several hyphen points can be placed in each word, 
whereas pronunciation indicates that the word cannot be hyphenated 
because it contains only one syllable. 

3. A number of English words also contain consecutive vowels that are 
pronounced as one vowel. The words THROUGH and BLEED contain 
examples of this condition. A computerized vowel scan can solve this 
problem by considering contiguous vowels as a single vowel. However, this 
assumption will produce incorrect results in a word like REALITY, where 
consecutive vowels are pronounced separately. 

4. The letter Y sometimes behaves as a consonant and sometimes as a vowel. 
In the word SYLLABLE, the Y behaves as a vowel, while in the word 
VINEYARD the Y behaves as a consonant. In general, if a Y is preceded by 
a vowel, it can be considered a consonant. 

The problems caused by variations in pronunciation are epitomized by the 
homographs which exist in the English language. Homographs are words 
that are spelled identically but pronounced differently and, therefore, 
possibly divided differently. For example, the word PRESENT can be either 
a noun or a verb depending upon the context in which it is used. If the word is 
used as a noun, it is hyphenated after the S; if it is used as a verb, it is 
hyphenated before the S. 

Composition Input 111 



Each of the problems noted above exists because syllabication is based upon 
pronunciation, and pronunciation is often unrelated to the alphabetic 
structure of the word. However, there are a number of words which are not 
hyphenated according to the way they are pronounced. For example, the 
words DOUBLE, MILITIA, and VISION are hyphenated according to 
conventions established by printers and writers. Word division is actually 
based on many different considerations: pronunciation, conventions and 
traditions in general use by printers and writers, etymology, component 
elements, context, etc. Considering the diverse bases for word division, it is 
understandable why so many inconsistencies exist in the hyphenation of 
English words. The inconsistencies resulting from just the variations in 
pronunciation should give some idea of the formidable problems that had to 
be solved in computerizing word division. 

Words which contain alphabetic and nonalphabetic characters or words 
which consist entirely of nonalphabetic characters are also problems: 

1. Alphameric words consisting of letters and numerals. This type of word 
most commonly occurs in parts catalogs and technical publications. 

2. Words that contain punctuation marks such as periods, question marks, 
and commas, when that word ends a line or a phrase. 

3. Compound words or phrases whose elements are separated by text- 
supplied hyphens or em dashes. For example, an em dash usually separates 
the elements of a telephone number, and a text-supplied hyphen usually 
separates the elements of a compound word such as RE-ENTER. 

4. Words that contain apostrophes; for example, the possessive form of a 
word (HENRY’S) or a contraction (WON’T). 

5. Words like “MacDonald” that contain both upper and lowercase 
alphabetic letters in the word. 

6. Words consisting entirely of nonalphabetic characters, for example, 
$12.38 and 1,000,000. 

The problem inherent in the above examples is not to recognize the 
nonalphabetic characters but to be able to make some decision about 
hyphenation-point placement based on these characters. For example, words 
which contain text-supplied hyphens can be broken at the point where the 
text-supplied hyphen occurs, and words like “MacDonald” can be divided 
just before the change from lowercase to uppercase characters. The problem 
of dividing mixed words is compounded by the fact that rules for mixed- 
word division are determined to a large extent by the stylistic inclinations of 
the user. Nevertheless, a generalized hyphenation routine must provide the 
ability to hyphenate mixed words. 

Not all of the problems associated with computerized hyphenation are due 
solely to the idiosyncracies of the English language or to the requirement to 
hyphenate mixed words. Two of the primary objectives of any hyphenation 
routine must be to maximize hyphenation accuracy and to minimize the time 
required to hyphenate the average word. The fastest and most accurate 
hyphenation routine might not be practical for the general user if a large and 
expensive system configuration were required. 

With the application of computers to text processing and the requirement for 
112 Composition Input 



the division of words at the end of a line, a number of techniques for the 
division of words have been developed. In general, these techniques can be 
divided into two basic types — dictionary and rules (algorihmic methods.) 

Perhaps the simplest computerized word division technique consists of 
storing all commonly used words with their associated hyphen points and 
then searching this huge dictionary each time a word is to be hyphenated. A 
primary disadvantage is that every word requiring hyphenation may not be 
in the dictionary. In order for the dictionary to contain all the words that 
might need to be hyphenated, especially when one considers proper names 
and “made up” words, computer storage requirements and execution time 
would be prohibitive. 

A second dictionary approach consists of storing commonly used suffixes 
and searching this suffix table to see whether the word to be hyphenated 
contains one of the suffixes. Since English is fundamentally an agglutinative 
language, a hyphen point usually occurs in the word before the suffix. Even 
though suffix analysis can pinpoint a word division point in some cases, a 
number of exceptions to these rules do exist. For example, the suffix -ING is 
usually preceded by a hyphen point, as in the words SAIL-ING, GO-ING, 
and COM-ING. However, some technique must be employed to correctly 
hyphenate words such as PE-KING, BET-TING, and FLEDG-LING, since 
these are exceptions to the general rule. In these cases, the suffixes -KING, - 
TING, and -LING can also be made suffixes and placed in a suffix table so 
that they will be searched before -ING. Even though a certain amount of 
hyphen placement data can be obtained from a suffix analysis, these 
examples point out the difficulty of compiling a proper suffix dictionary in 
order to minimize the number of exceptions. 

A prefix table may also be utilized for the same reasons that a suffix table is 
used. But the same problem of exceptions also exists with prefixes. For 
example, the prefix TRANS- appears in words like TRANS-PARENT and 
TRANS-FORM, but the words TRAN-SCRIPT and TRAN-SCEND 
violate the rule. An analysis of a large sample of English language words 
indicates that there are a great many more exceptions to prefix analysis than 
there are to suffix analysis; for this reason, prefix analysis is difficult to 
implement, and quite often the amount of coding needed is not worth the 
amount of hyphen placement data obtained. 

The second general type of computerized hyphenation technique is generally 
referred to as rules or algorithmic. An algorithm is a well defined process or 
set of rules for the solution of a problem in a given number of steps. With 
relation to hyphenation, the rules are usually the result of extensive 
statistical or phonetic and contextual analysis of a large sample of English 
words. The development of any algorithm is plagued by two major 
difficulties: 

1. The accuracy of the algorithm is directly related to the content and size of 
the data base from which it is derived. An algorithm based on a small sample 
of words which does not include the majority of commonly used words 
cannot be expected to correctly hyphenate a large portion of commonly used 
words. An algorithm derived from a data base of only scientific terms, for 


Composition Input 113 



example, may be highly inaccurate when used to hyphenate proper names or 
words from another discipline. 

2. Normally, most computerized algorithms can analyze a word for division 
points based only on the word’s alphabetic structure. Hyphenation of 
English words is primarily based on phonetic or contextual considerations; 
however, each algorithm must in some way solve this problem if a 
reasonable amount of hyphenation accuracy is to be obtained. 

Because of the two general difficulties noted above, most of the currently 
used algorithms have one major shortcoming —the existence of exceptions 
to the rules of the algorithm. 

One algorithm that has been used is based on the observation that, for a 
greater number of words, hyphen points exist after the third, fifth, and 
seventh letters of the word. This algorithm is surprisingly accurate 
considering its simplicity, but the number of exceptions to this rule are too 
many to permit its use without extensive augmentation with other 
techniques. 

One algorithmic technique consists of storing, for each of a large number of 
strings of characters, the probability that that string of characters comprises 
a syllable. 

Another algorithmic method that has been developed is commonly known as 
the tree technique. Starting at the first vowel of the word, each successive 
character in the word is analyzed to see whether this character, in 
conjunction with the preceding characters, indicates anything definite about 
hyphen placement in the word. 

If the character following an O is other than an M, the associated processing 
routine may indicate that a third character needs to be analyzed. This 
branching process continues until the end of the word is reached, at which 
time all the hyphen points in the word will have been found. This technique 
is, in effect, a type of dictionary lookup that attempts to pinpoint common 
characteristics of character strings without looking at the entire word. If this 
technique were flowcharted, the result would be very similar to a tree with its 
many branches. 

An important algorithmic method of hyphenation is based on the analysis of 
the probability that a hyphen point will occur before, between, or after 
consecutive two-character combinations comprising the word. Using a large 
correctly hyphenated dictionary as the data base, the probabilities are 
determined by dividing the total number of times that the two-letter 
combinations occur in all the words with a hyphen point between them by 
the total number of times that the two-letter combinations occur in the word 
regardless of whether a hyphen point separates them or not. The 
probabilities themselves are stored in large tables. The algorithm steps 
across the word two characters at a time, using the probability for each 
possible hyphen point in the word. 

A more detailed description of this method may be helpful at this point. 
Consider the word PROGRAM. The algorithm initially considers this word 


114 Composition Input 



to have six positions at which division points can occur: 

P-R-O-G-R-A-M 
1 2 3 4 5 6 

Positions 1, 2, and 6 can be eliminated from consideration, since each 
syllable of the word must have at least one vowel in it, and if a hyphen were 
placed at any of these points, this condition would be violated. Hence, the 
analysis is limited to positions 3, 4, and 5. 


Hyphenation accuracy and its effects. 

The goal of composition people today seems to be 99 percent accuracy. On 
long-measure book work with today’s programs, preliminary experience has 
indicated computers will hyphenate between one line in 100 and one line in 
1,000. The former hyphenation rate coupled with even 99 percent acceptable 
accuracy would require resetting of two lines in 10,000. In a 400 page book, 
40 lines per page, this would amount to resetting two to four lines. 

The various approaches being taken today seem to be converging toward a 
combination system - a basic logic program supplemented by a small 
exception dictionary of words that cannot be broken properly by the rules. 
The dictionary area of the computer’s memory would also have space 
provided for adding or revising words as the language changes. 

Opinions differ, however, on how large the exception dictionary will have to 
be to achieve this goal. One manufacturer believes it can be reached with 
2,000 words, while another maintains it requires 12,000 words. The final 
answer depends on several factors, such as the extent to which logic 
programs can be perfected and the varying needs of different composing 
rooms. 

These varying needs are of special significance to book manufacturers today. 
Even with computer programmers specifying Webster’s first choice with 
their 99 percent accuracy figures, other typographic standards may not be 
met adequately by these programs. Most of the work done to date has been 
done in newspapers; computer typesetting is only beginning on a large scale 
in book manufacturing. And newspaper quality requirements differ greatly 
from book requirements. 

For instance, newspapers hyphenate proper names, a practice that may not 
be permissible in some books. The computer could be given a rule to avoid 
this - such as don’t hyphenate words beginning with a capital letter - but this 
again takes more space in the computer’s memory as well as slowing down 
operations. 

Another instance is the breaking of words containing ligatures. Provision for 
hyphenating between characters forming ligatures is not included in many 
programs currently in use. Programming this requirement is more complex 
than a first glance would indicate. 


Composition Input 115 



FIVE-MINUTE TYPING TEST 


Without a doubt very few of us on entering the average modern business 
establishment have ever stopped to consider the vital role played by records 

in the successful operation of that business. To say that business as we 

* 

know it would not be possible without the use of records is certainly no 
overstatement. 

To say that we live in a world of records is a self-evident truth. 

Our lives are bounded by them from the cradle to the grave. The progress of 
civilization in the various countries of the world can be measured by the care 
and completeness with which the experiences and discoveries of one generation 
have been passed on to the next. Business records, naturally, have developed 
to the highest degree in centers of commercial activities. 

To the uninformed, "Keeping books" may be regarded as an activity of 
recent origin and as being responsible for many of the hidden mysteries of 
the business world. On the contrary, however, records are almost as old as 
history itself. The crude characters scratched on the walls of the caves in 
which primitive man existed, the baked clay tablets in the Near East, or the 
papyrus rolls in Egypt tell us that since earliest times man has been confronted 
with the problems of recording what belongs to him, what is due him, and what 
he owes to his creditors in order that he may have some idea of his financial 
progress and status. In fact, only two years after Christopher Columbus made 
his way westward across the uncharted Atlantic Ocean, the first textbook in 
bookkeeping was written by an Italian monk. Those principles of keeping 
accounts by the double-entry system as developed in his book are basic to all 
our modern accounting records. 


Stroke s 

71 

148 

222 

295 

311 

378 

455 

534 

612 

690 

750 

819 

894 

970 

1047 

1124 

1205 

1283 

1361 

1439 

1515 

1588 

1667 

1697 


116 Composition Input 



9. Word 


processing 


This section is included because it will probably require a book of its own 
someday. Word Processing is the new name for what we once knew as 
automatic typewriting. Almost daily, its application to the composition 
function becomes more defined. 

If there is any trend that is developing in the composition input area it is 
that of “capturing keystrokes.” at the point of origin. The concept is 
elementary: at the time someone starts to put an idea on paper capture 
some element of that process which can be used as input for typesetting. 
The word processor, whatever make, allows ideas to be typed at draft 
speeds without regard to errors. Corrections are made only in those areas 
that require correction. Thus a typist can produce manuscript copy more 
rapidly, and in addition to the typescript maintain some kind of record of 
that typing. Magnetic tape cassettes are the most popular form of record 
keeping since they can be used over and over again. 

The point at which I first write the paragraph is the source documentation 
stage. I could be a reporter or an author or anyone who has ideas to 
generate and ultimately communicate. I would type this information as 
quickly as possible or dictate to a secretary who would do the same. Next 
I would go back and change or edit my original material (or the secretary 
would). I can change or insert new data because the machine has within it 
a flexible recording medium that permits this and the logic to allow it to 
be used effectively. 

The advantages to anyone who sets type: less redundant keystrokes (don’t 
forget that all typesetting material is actually keyed twice, once by the 
originator and then by the typesetter), easier correction of author’s 
alterations and simple editing. All this is done without a computer. This 
chapter will attempt to deal with some of the concepts involved in word 
processing and some of the developments that could affect composition 
input. 


Composition Input 117 



Although our ultimate concern is with composition input this area is more 
clearly understood by first reviewing the concepts, techniques, and 
procedures used to create and prepare material for typesetting. 

In essence, one takes an idea and through a series of steps, produces multiple 
copies in the form of books, newspapers, or other printed matter. 

Creation of material 

No matter what the material is, be it a letter, a textbook, a newspaper 
article, a magazine, or an advertisement, the starting point is the same: an 
idea in the mind of a person. To take this basic idea and create a useful 
document or manuscript involves two fundamental functions: the creative 
function, and the corrective function. 

To start the cycle, an author creates a document containing his ideas. This 
may be handwritten; it may be typed in rough draft form; or it may even be 
in the form of dictation. The idea then passes to the corrective stage, where 
the idea is improved. The form of the improvement can be varied. It may be 
a simple spelling correction, it may be improvement of the grammar, or it 
may involve massive re-organization of the material with respect to sequence 
or content. The improvements are then sent back to the author for his 
reaction. The author may massage the material further and then pass it back 
to the corrective function for further improvement. It is entirely possible for 
material to go through this inter-active “loop” many times before the 
material is acceptable. It is also possible for the creative function (author) 
and the corrective function (editor) to be the same person. When both 
creative and corrective functions agree that the material is acceptable, it is 
then considered to be a good manuscript. 

Depending upon the remainder of the system, the manuscript may be in one 
of two forms: (a) it can be a stack of typewritten pages, or (b) it can be in 
machine readable form such as magnetic or paper tape. 

Processing of the material 

Once the manuscript is informationally correct and in its final form, it 
crosses the line from “art” and moves into the field of “craft” where it 
begins the printing process. 

Typewriting 

The creation of a page eventually involves the use of a typewriter to put the 
authors’ ideas on paper. Usually this is done at the very start, and the 
enrichment cycles involve a re-typing operation to incorporate the changes. 
Of course, any typewriter can be used to create correct pages, but our 
interest lies with those machines which have been designed to automate and 
improve the process. 

Step 1 

The manuscript, depending upon its form, is given to either a keyboard 
operator or a converter operator. 


118 Composition Input 



a) The keyboard operator simply keys the information onto punched paper 
tape. This tape contains both the textual content of the manuscript, plus the 
appropriate typesetter control commands for font style, indents, etc. 

b) The converter operator loads the machine readable record on the input 
side of the converter, and produces a new tape that is capable of driving the 
typesetter. This tape contains both textual information, plus appropriate 
typesetter commands to control font style, indents, etc. 

Step 2 

The typesetting process takes the input tape and produces galleys. While the 
input tape contains the textual information and some typesetter commands 
such as font style and indents, the typesetter is typically adjusted to produce 
the desired line length, leading, etc. 

Step 3 

The make-up step takes the galleys produced by the typesetter and arranges 
this information to form a full page. In addition, all artwork, drawings, 
illustrations, etc, are added at this step. 

Step 4 

The full page, called “camera ready copy”, is now passed on to the camera 
stage. A photograph is taken of the page, and a negative is made. In full 
color printing, a color separation process is used, and the original camera 
ready copy is photographed 4 times. A separate negative is made for each of 
the 4 colors to be printed, yellow, magenta, cyan, and black. 

Step 5 

The megative (or negatives) now go to a platemaking process, where 
photographically treated aluminum sheet is used to make a contact print of 
the negative. In full color printing 4 such aluminum contact prints would be 
made, one for each color used. This aluminum sheet is now the offset plate 
master. 

Step 6 

The offset plates are now loaded on the press, and the printing begins. In 
black and white printing, the paper need pass through the press only once. In 
color printing, the paper needs to pass through the press once for each color, 
or more practically, pass through a printing “head” for each color. Thus, full 
color offset presses are sometimes referred to as 2 head or 4 head presses. 
The output of the press is the finished product, and all that needs to be done 
is the appropriate folding, collating, stapling, or binding. 

This is one example of the loop that begins with an idea and ends in such a 
way as to transmit that idea to others. 

Basic automatic typewriters 

Virtually every automatic typing system revolves around the use of an 


Composition Input 119 



electric office typewriter that has been equipped with the facility to accept 
electrical signals from the outside to activate typing, and to generate 
electrical signals as a by-product of typing. While the form of this capability 
varies greatly with manufacturer, these machines are commonly referred to 
as I/O (Input/Output) typewriters. The I/O typewriter forms the main 
element of the system, and to the typewriter is added a recorder (usually a 
slow speed punched paper or magnetic tape device), a reader (again either 
magnetic tape or punched paper tape) and the necessary control electronics. 
In principle, the typewriter will create a machine readable record as a by¬ 
product of typing, and it can read this record and produce page copy. Some 
systems can read tape, type, and produce new tape simultaneously, while 
other systems can only read and type, or type and record. 

A number of different typewriters have been equipped with I/O features and 
used in automatic typing systems, but the most common machine for this 
application is the IBM Selectric, or Golf Ball typewriter. This machine 
operates at a printing speed of 15.5 characters per second, and offers the 
ability to quickly change character sets and/or styles by simply changing the 
ball. A number of type bar machines have also been offered with I/O 
features. Type bar machines usually operate at speeds of around 12 
characters per second. While there are a number of differences between the 
Selectric typewriter and the type bar machines, the two essential ones are 
speed and code flexibility. The Selectric is a faster machine, but it has a 
rather rigid code structure. The type bar machines, while slower, offer far 
greater flexibility for utilizing different code sets. 

Typewriter speeds are usually quoted in terms of characters per second, 
while typing speeds are quoted in terms of words per minute. Keyboarding 
speeds are quoted in characters (or keystrokes) per hour, while Keyboarding 
of newspaper material is quoted in terms of lines per hour. The table 
furnishes a quick cross reference to these different standards. By definition, a 
word is 5 characters plus space, or 6 characters. A newspaper line is 
commonly 30 characters. 

The concept of power typing 

Power typing is a term frequently used by IBM, and in principle is based on 
the fact that a girl can type faster when she does not have to worry about 
creating error-free page copy. A typical rough draft speed of a typist is 50-60 
words per minute when trying to type error-free copy. Power typing involves 
giving this girl an automatic typing system that creates a machine readable 
record as a by-product of her rough draft typing. Further, the system permits 
the girl to back up and overstrike any typing errors she may make, and insert 
the correct information. This procedure also backs up the output tape, 
deleting the erroneous information and inserting the correct material. It is 
important to note that the desired output of this kind of system is a good 
tape ; the page produced on the typewriter serves only to give the operator 
visibility as she types, and has no further value. 

Once the tape has been produced, it can be handled one of two ways. 

Stand alone system 

In this case, the same girl that prepared the tape in a rough draft typing 
120 Composition Input 


Electric Typewriters 

Adler 

Facit 

Hermes 

IBM 

Olivetti 

Olympia 

Remington 

Royal 

SCM 

Automatic Typewriters 

American Automatic Typewriter Co. 

Auto-Typist 

Metro-tel 

Mate 

“Automatic” Text Editing Devices 
(Stand-alone Work Stations) 

IBM 

MT/ST-II & IV 
MC/ST 

Itel 

815 

852 

853 

Ty-Data 

3600/1 

3600/2 

3600/3 

QuinnData 
Quinn Type-70 
Quinn Type-80 



Edityper 

200 

Redactron 

1 card 

2 cards 

1 cassette 

2 cassettes 

Remington Rand 

1 card 

2 cards 

1 cassette 

2 cassettes 

Wang Laboratories 
1210 
1220 

Multiple Work Station 
Dedicated Computer Systems 
Information Control Systems Inc. 
Astrocomp 
Edit Systems Inc. 

Text-Ed 

Time Shared Text Editing 
Bowne Time Sharing 
Word I One 

Other time sharing companies in U.S. 

CRT Stand Alone Devices 
Spiras Division of USM Corp. 

A ecu text 

Lexitron Corporation 
Video type 


mode now re-inserts this tape into her typing system. The tape now produces 
clean page copy at the maximum speed of the typewriter. An example of this 
kind of process is as follows: suppose a girl were to type a 10,000 word 
document. By using rough draft typing, it will require 250 minutes (at 40 
words per minute) to create a good tape of this information. The tape is then 
played back through the typing system, and it “power types” at a speed of 
155 words per minute (IBM MT/ST), which takes an add minutes. The 
entire task has consumed 315 typing minutes, wheras the same task, done on 
a standard typewriter by the same girl, would require 667 minutes of typing 
time, at an average typing speed of 15 words per minute. 

The basic concept of the stand alone system is to use the same machine for 
both rough draft typing and power typing. 

Centralized typing system 

A slightly different set of circumstances surrounds the central typing 
application, where there may be many girls transcribing information from a 
central dictation source. These systems are commonly used by hospitals and 
large insurance offices. The first step in the process is for a girl to transcribe 
material in a rough draft typing mode. The good tape that is created by this 
process then migrates to a central playback system, where the tapes are 
power typed at a speed of 155 words per minute. In this situation, three girls 
power typing and one girl operating the playback system can produce the 
same amount of finished copy as 7 girls using standard typewriters. The 
basic concept of the centralized system is the use of different machines for 
each task. 

The use of a complete automatic typing system to produce good tapes in the 
centralized system is hard to justify, since the desired end result is a good 
tape and page is of no value, Opportunities exist for the application of a 
lower cost keyboard recorder, which produced good tape as the primary 
output but does not deny the operator visibility of her work. 

Word processing automatic typewriters 

Automatic typewriter systems used for the creating and updating of typed 
pages perform the following functions: 

- create a machine readable record as a by-product of typing. 

- permit re-creation of the correct portions of the copy by using this 
machine readable record to operate the typewriter 

- permit corrections to be introduced in the proper place in the typed 
material, producing new page copy 

- create a new machine readable record, containing the useful portions of 
the original material plus the additions. 

There have been a number of companies that have produced automatic 
typing systems, including Facit, Tally, Smith Corona, and others. Of the lot, 
IBM, Dura and Friden survive as the largest suppliers of word processing 


Composition Input 121 



automatic typewriters. These three systems will be exposed in the following 
material. 

Word processing typing systems are essentially automatic typing systems 
with some additional control features provided to simplify and improve the 
editing/revision tasks. The required control features are three: 


1) Automatic carriage return. This is accomplished by using a “hot zone” 
right margin feature. The hot zone is determined by the setting of the right 
margin stop, and once the typewriter has typed to the right margin (or hot 
zone) a space code or hyphen code causes the carriage to return. This 
permits the operator to do some formatting of the typed copy by 
determining the line length of the finished page. The machine ignores any 
carriage return codes in the input tape, which eliminates the possibility of 
format problems due to carriage return codes being discarded with unwanted 
material in the input tape. 

2) Discretionary and mandatory hyphens. Two kinds of hyphens are used in 
word processing typing systems. Discretionary hyphens are those that print 
only when they appear in the hot zone. As the original material is typed, the 
tape may contain a number of words that are broken by using discrectionary 
hyphens. When the page copy is reformatted, these words may not appear at 
the end of a line, and so the hyphen is redundant and is not printed. 
However, when one of these hyphens appears in the hot zone, it is printed, 
and it initiates a carriage return. Mandatory hyphens are those that print 
every time, regardless of their place in the line. Some word sequences, such 
as “mother-in-law”, require hyphens to always be present. When a 
mandatory hyphen appears in the hot zone, it also initiates a carriage return. 
The two different hyphens carry two distinct codes; and word processing 
typing systems have two hyphen keys on the keyboard. 


3) Programmed playback and stop. This feature permits the operator to 
select a certain amount of information to type, after which the machine will 
stop. Most machines offer some combination of character, word, line, 
sentence, or paragraph. In this way, the operator can instruct the machine to 
type only a specific portion of the material, which permits additions or 
corrections to be inserted in the desired spot. 

The logic element controls the operation of the entire system, and this 
operation is based on commands from the control panel, or even from 
command codes that appear in the input tape. For example, when two 
readers are being used in a “tape merge” mode, transfer codes in the input 
tapes will shift the input from one reader to the other. Other codes that can 
be in the input tape are “punch off’ (where the material is to be typed, but 
not copied), etc. The word processor element determines how much material 
will be printed before the machine stops (character, word, sentence, etc.). 
The example given above uses paper tape terminology, but the same 
principles apply to magnetic tape systems. 

IBM MT'/ST (Magnetic Tape Selectric Typewriter) 

This is the most well known of all the automatic typing systems. The heart of 


122 Composition Input 



the system is the IBM Selectric typewriter, to which is attached either one or 
two magnetic tape stations. Each station is capable of either reading 
information into the typewriter or recording material from the typewriter. 
The logic in the machine permits the playback to stop after each character, 
word, or line. The logic also searches for a given line number in the tape, and 
the search speed is 900 characters per second. In addition, the two station 
configuration can “merge” tapes, which combines portions of two input 
tapes to form a page copy. One of the tapes can serve as a program tape to 
determine format, and the other tape serves as the material input tape. The 
magnetic cartridges use special edge-sprocketed magnetic tape, very similar 
in appearance to 16 mm motion picture film. The cartridges hold up to 100 
feet of tape, which will store 24,000 characters, at a price of about $20 each. 
Cartridges with less tape are available for $12 each. The cartridges are self 
threading, and re-usable many times. 

IBM MC/ST (Magnetic Card Selectric Typewriter) 

A recent addition to the IBM automatic typing system line is the MC/ST. 
While essentially the same as the MT/ST in principle, it differs in one 
important aspect. The media used is a special magnetic card, which can store 
5000 characters arranged in 50 tracks of 100 characters each. Each card 
typically contains a full page of typed information. The MC/ST uses a single 
record/replay station, and handles a single card at a time. The card is re¬ 
usable many times, and cards cost about $1 each. 

The significant difference between the MT/ST and the MC/ST is the 
concept of updating. The MT/ST reads an input tape, and creates a new 
output tape which contains all the valid material from the original tape plus 
any additions or corrections. The MC/ST reads the input card, and any 
changes or additions are made on that same card, thus in effect updating the 
original card. It’s obvious that the MC/ST is therefore limited in the 
amount of new information that can be added, while the MT/ST can accept 
virtually any amount of new material. The MT/ST is basically an editing 
system, while the MC/ST is aimed at the concept of single station power 
typing. 

Dura 

The punched tape equivalent of the MT/ST is the Dura 1041 equipped with 
word processor. The heart of the machine is the IBM Selectric I/O 
typewriter, to which is connected a slow speed paper tape reader and punch. 
The word processor feature permits the playback to occur a word, sentence, 
or paragraph at a time. In addition, a single step key furnishes the character 
at a time playback. The concept is to read an input tape, insert corrections or 
additional information, and create a new output tape as a result. The 
machine can also be equipped with an additional reader, which permits 
merging of information from two tapes. An additional option permits the 
use of a slave printer, which then makes a powerful system for the mass 
preparation of personalized letters. The primary input reader contains the 
letter text, and the auxiliary reader contains all the names and address. The 
main printer types the letter, and the auxiliary printer types the mailing 
envelope at the same time the inside address is typed on the letter. In 
addition to handling punched paper tape, the machine can be equipped to 
handle edge punched cards, either singly or in fan fold strips. 

Composition Input 123 



The Friden machine has been around for a number of years, having started 
out in life as an I/O device for small computers. The heart of the system is 
the Friden type bar I/O typewriter, which was originally designed and 
manufactured by IBM as their model A typewriter. The typewriter is used in 
conjunction with a Friden slow speed paper tape punch and reader, and in 
concept is virtually identical to the Dura 1041. The word processer of the 
Flexowriter is based on playback of a character, word, or sentence at a time. 
Like Dura, the Friden can be equipped with a second reader, which permits 
some merging-of input information. The Friden also handles edge punched 
cards. 

4) Computer managed typing systems 

A number of companies are now offering typing systems which address the 
very same problem as the word processing systems such as the MT/ST; 
however, they utilize a minicomputer as the basic control element. Most of 
these systems are designed to furnish a good typed page as the primary 
output, and options are available to furnish either paper or magnetic tape as 
well. The computer, in addition to controlling the typing of the manuscript 
information, can also be used to perform “pre-processing” of the material, 
thus organizing the typed material so that the resultant output tape can be 
fed directly into a data processing or typesetting system. 

A computer managed system is the ASTROTYPE, offered by Information 
Control Systems. 

The system consists of up to four Selectric I/O typewriters, connected to the 
computer. Up to four DEC-TAPE magnetic tape units are connected to the 
output of the computer, in effect giving one tape unit for each typewriter, 
under computer control. In use, the operator types material at rough draft 
speeds (power typing), and the information is captured on magnetic tape. 
The computer automatically assigns line numbers to each line as the 
operator types. The page copy then passes on to the edit function, where 
appropriate notations are made relative to the portions of the text that must 
be changed. The page copy comes back to the operator, and now she merely 
loads the matching tape, and uses the typewriter to make statements 
concerning the action to be taken. The operator types in the number of the 
line in question, and then tells the computer what must be done (remove a 
word, change a word, change a letter). The computer makes the requested 
change, and types out the line for the operator to visually verify. The 
operator must acknowledge that the revision is correct, and it then becomes 
part of the mag tape record. When all the corrections, additions, or revisions 
have been done, the operator calls for a print out of the Final page. The 
computer then generates the page copy at the rate of 155 words per minute. 
The magnetic tape record of the page is stored until the next revision is 
required, or until the tape is erased. 

The computer can also do some formatting of the page, such as changing line 
lengths, and justifying both margins by interword spacing. The computer can 
also remove a paragraph from one section of the page and re-insert it 
elsewhere on the page, or it can be inserted on another page. 

The system is available in a number of configurations. It can be equipped 


124 Composition Input 



with a line printer, or it can also be furnished with an output of either 
magnetic or punched paper tape. The basic system consists of the computer 
(DEC PDP-8), and one typing terminal. To this may be added typing 
terminals (up to a maximum of 4 terminals). The system also offers the line 
printer. 

A somewhat different system is the Varitext, offered by Varian. 

The system can be used with up to eight typing stations. The typing station is 
an IBM Selectric typewriter, modified by Varian to hold two tape cassettes 
of the Philips/Norelco style. In operation, the operator uses the KSR 33 
Teletype machine to instruct the system as to the task to be done. Then, the 
typing station operators type information in a power typing mode, recording 
the information on the cassette in the typewriter. When the page (or pages) 
have been completed, the typing station reads this cassette to the computer, 
and the computer assigns line numbers and prints the information on a high 
speed line printer. The line printer output is then passed to the edit function, 
where correctional notations are made relative to the changes required. The 
page then returns to the typist, who finds the matching cassette and loads it 
in the typewriter, along with a blank cassette. The typist states the line 
number in question, and the computer controls the cassettes in the terminal 
to copy all the good information from the original cassette to the new one, 
stopping at the desired line. The typist enters the correct information, which 
is recorded on the second cassette. The typist states the next line, and the 
process is repeated until all corrections have been made. When the new 
cassette is completed, it can be read to the computer for a new page print out 
(either with or without line numbers) or the same cassette can drive the 
typing terminal to generate page copy. This system also offers either an IBM 
compatible mag tape output, or a high speed paper tape punch. 

A still different system is one using a time sharing terminal, such as the IBM 
2741 or the Datel. These are typing terminals that can be connected to a 
large central computer by means of telephone lines. The concept was 
originated by IBM, woo offered their ATS (Administrative Terminal 
System for data text) through their Service Bureau Corporation. Several 
companies offer versions of the service, among them VIP Systems in 
Washington DC. 

The typing terminal is an IBM Selectric I/O Typewriter with some 
electronics. When not connected to the computer, it operates just like an 
office typewriter. In use, the operator connects the terminal to the computer 
by daking a telephone call. As information is being typed, it is also being 
stored at the computer. The computer automatically numbers lines. Thus, 
the page copy is with the operator and the stored information is with the 
computer. The page copy is passed to the edit function, and notations 
relative to corrections are made. This page then comes back to the operator, 
who re-connects the terminal to the computer, types in the in the address of 
the information so the computer can retreive it from memory, and then 
states the corrections that must be made by using line numbers to pinpoint 
areas of interest. The computer makes the corrections, and when all 
corrections are complete, the operator can call for a print out of the updated 
information. Since the central computer is usually large (such as the IBM 
360/65) it is possible for a fully formatted magnetic tape to be made of the 
material. 


Composition Input 125 



Typing terminals are available from a number of sources, and Dura makes a 
version of the 1041 typing system for this use. In the case of Dura, the 
machine is furnished with a paper tape reader and punch, so that in addition 
to the operator having page copy she also retains a paper tape record of the 
material. 

The rate structure for time sharing varies with the company. In general, the 
charges are some combination of the following. The prices shown are those 
quoted by one service bureau. 

Typing terminal model 2741 $100 per month 

Connect time $ 11 per hour 

CPU time 150 per second 

The CPU time rate is usually governed by the size of the core resident in the 
computer. The larger the core size, the higher the rate. In addition, it is 
possible to leave information stored with the central computer. 

The systems presented in this section have all been based on the principle 
that the primary output is a good page ; the good tape is secondary. There 
are a few systems (such as the Astrocomp) that are based on the converse, 
that is, the good tape is the primary output, the page is secondary. There are 
other systems that take this point a step further, and provide a good tape but 
no printed output. As a general rule, these systems do not live in both the 
world of creation and the world of typesetting. 

The typesetting function 

The typesetting function embraces the combination of two different types of 
information 

the text.... what does it say? 
the format... how does it look? 

Virtually every typesetting device requires that the format control 
information be interspersed with the text material. For our purposes, all 
typesetting tasks will be placed in one of the three following catagories, 
depending upon the amount of format control information required 

1) Straight matter or text... doesn’t require much format control 

2) Display/advertising material... takes a fair amount of format control 

3) Textbooks and special problems ... takes a lot of format control 

It’s obvious that the actual procedure in any given case will be determined by 
such things as 

— the type, and capability of, the typesetting system being used 

— the form of the input data 

— the output desired 


126 Composition Input 



Straight matter or text 


Straight matter is best illustrated by using a newspaper or paper back book 
as an example. The task is to typeset large volumes of material, and the 
material is simple in format and uses only a few basic fonts. Straight matter 
typesetters usually contain two magazines (each with two rail choices) and 
the magazines are loaded with different size fonts, such as 6 point and 10 
point. The type styles used are fairly common, such as Bodoni, Garamond, 
or News Gothic. The rail choices furnish bold face or italic. The format is 
generally consistent, and doesn’t change often. Newspapers use a standard 
11 pica line for news columns, with front page or editorial columns going to 
15 or 20 pica line lengths. Paper back books are usually around 22 pica line 
lengths. 

Straight matter tasks can be handled by OCR, or by direct reading of tapes 
produced by Dura or Friden or other machines. However, two things must 
be done to satisfy the requirements of the typesetting machine: 

1) The codes must usually be converted, since computer codes are not 
normally compatible with most straight matter typesetters. 

2) The textual matter must be combined with typesetter commands to 
achieve formatting of the material. 


OCR systems usually require two passes at the material before it is ready to 
go to the typesetter. The first pass is interpretive, where the pages are read 
and converted to machine readable tape, with the output being either 7 or 9 
track IBM magnetic tape format, or 8 level punched paper tape. 
OCRsystems can read special character sequences which are interpreted as 
command codes. Since only a limited number of characters exist on a 
typewriter, it is necessary to combine codes to create command sequences.In 
this way, the “$” followed by some non-digit such as a letter or a 
punctuation mark can be inserted on the page copy to stand for Quad center, 
or Upper Rail. During the interpretive pass, this code sequence is simply 
converted to a discrete 8 bit code. The second pass is the translative 
operation, where the computer compatible tape is read into the system and a 
typesetter format tape is generated. Most straight matter typesetters require 
6 level TTS tape for input, and in addition, command sequences frequently 
take 2 or more codes. The translative pass generates the proper 6 level code 
for the textual characters, and “explodes” the 8 level single character into 
the appropriate string of 6 level characters for functions. Some of the larger 
OCRsystems have sufficient computer capacity to accomplish both the 
interpretive and translative functions in a single pass, however the smaller 
machines based on the PDP-8 or the Varian generally require two passes. 

Classified ads 

Newspapers, of course, have requirements in excess of simple straight 
matter problems. Such things as classified advertising and display matter are 
also routinely done, however in virtually all cases, the tapes to drive the 
typesetters are produced on keyboards (as opposed to any OCR type 
system). The keyboard produced tapes are non-justified, however there is 
seldom any implication of line measure, since any line keyed is less than the 


Composition Input 127 



line measure being used. For example, most classified ad columns are 9 
picas, and provided the material keyed does not exceed 9 picas on any line, 
there is no line end decision necessary. 

Display typesetting 

Newspapers, as well as magazines, use advertising and display copy 
throughout their publications. The task is to create eyecatching copy, 
arranged in a well balanced aesthetically pleasing manner. This kind of 
material combines a variety of different font sizes, font styles, line lenghts, 
photographs, and special logos. 

Essentially, there are two ways to accomplish this kind of composition. 

Paste up 

The most common traditional method. In this procedure, all the elements of 
the material are typeset separately. For instance, all the 5 point material is 
keyed and typeset, then all the 8 point material is keyed and typeset, and so 
on. When all the pieces are complete, the ad is assembled in a make up 
function. The various textual elements are cut and pasted in their proper 
places and at the same time illustrations and photographs are inserted (in the 
form of halftones). Large type is usually typeset on a display or headline 
machine, such as the CG 7200 or the VGC Phototypositor. 

Compose 

This procedure involves the use of a phototypesetter (or more properly, 
photocomposer) that can typeset a wide variety of font sizes, and styles on a 
single piece of film or paper. A number of machines are available today, 
starting with expanded machines such as the CG ACM9000 and ending with 
the high speed high cost CRT systems. Machines such as the Merganthaler 
Super-Quick can hold 8 fonts in sizes from 5 to 72 points. The Photon 713-5 
holds 4 fonts, in sizes from 5 to 36 points. Some machines of this class offer 
such features as “reverse leading”, where the film or paper can be 
“advanced” in either direction to shift the baseline position of the desired 
character. Obviously, these additional features greatly complicate the 
keyboarding task, since the typesetter control is much more complex. As an 
example, a straight matter machine may require 1 or 2 characters to shift 
from one font to the other (upper magazine to lower magazine), while a font 
change command on a photocomposer may require a code sequence of up to 
10 characters. Photocomposers usually require justified tape as an input, 
which further complicates the keying task when changing font sizes. 

The choice of method (i.e., keyboard and typeset vs paste-up) is usually 
made by the mark-up or layout man. To adequately instruct the keyboard 
operator he must write down each and every step of the keying process 
necessary. Obviously, if the copy to be set is very complex, it may take him 
longer to figure out the keyboarding procedure than it would take to 
compose the material using cut and paste techniques. On the other hand, he 
may know that a certain keyboard operator is very skilled at that particular 
kind of keyboarding task, so he will make the decision to have it 
keyboarded. The actual decision is subjective, and is influenced by a number 


128 Composition Input 



of factors that are subject to change, such as the set up of the typesetter at 
that moment, the number of skilled keyboard operators available, et al. 

In general, there are three ways to create the tape required to drive these 
photocomposers: 

Counting keyboards 

In this method, a counting keyboard is used. The keyboard is adjusted for 
the desired line length, and an indicator on the keyboard tells the operator 
how much space is left in the line as the keying is done. Where font changes 
are necessary, the keyboard operator uses a “look up table” that furnishes 
the character sequence necessary to do the task. These counting keyboards 
usually have 2 or more width plugs “on line”, to facilitate the use of fonts 
with different width values. Counting values can be selected by a single 
keystroke. These keyboards are available in either blind or hard copy 
versions. 

Stored format keyboard 

The Automix 710 machine is the primary example of a counting keyboard 
with format storage. The 710 is a typewriter based system, which contains 6 
programmable “magazines”, and each magazine can be independently 
adjusted for a full set of criteria: font size , font style, line length, leading, 
space band value, etc.. The operator then uses a single keystroke, calls up the 
command sequence and inserts it in the output tape automatically. With this 
system, the operator need concentrate only the text, since all the complex 
routines are electronically generated. 

Computer systems 

These systems are extensions of the basic hyphenation-justification 
computers used with straight matter machines. The role of the computer has 
expanded to embrace photocomposer commands, in addition to the simple 
line measure task. In a computer system, the photocomposer control 
sequences are resident in memory. At the time of keyboarding, the keyboard 
operator would insert an address string of up to 9 characters in the tape. 
When this unjustified tape was read into the computer, the address string 
would cause the computer to insert the entire photocomposer control 
sequence in the resultant output tape. Since the composer command 
sequence could be as many as 20 characters, this saved the keyboard 
operator a significant number of keystrokes, as well as reducing the 
likelihood of errors. These also meant that virtually any keyboard, be it 
blind or hard copy, counting or non-counting could be used to prepare input 
tapes for computer systems. All that was necessary was the appropriate look 
table to inform the keyboard operator of the address string required. 

This computer implication did give rise to a family of keyboards that are 
specifically tailored for computer input. These keyboards offer an auxilary 
keyboard that generates a number of fixed code strings, or addresses, that 
can be called by a single keystroke. The code strings are usually factory 
wired, and as such are for use with a specific computer, such as the IBM 
1130 or the PDP-8 These keyboards are available in either blind or hard 
copy versions. 


Composition Input 129 



Textbooks and special problems 

The most complex kind of typesetting usually revolves around the 
preparation of textbooks, such as mathematics or languages. 

This type of work uses a wide number of special symbols, in addition to both 
light and bold face, and italic, typefaces. In addition, material can appear on 
different baselines, for both superscript and subscript notation. Typesetters 
used in text material composition are usually selected for their ability to 

— furnish adequate symbols 

— furnish adequate fonts 

— furnish adequate sizes of both symbols and fonts 

— be able to use multiple baselines 

Symbols 

The most important aspect. The special symbols are usually grouped 
together on “pi” grids on phototypesetters, and on a special “pi” ball for 
MT/SC systems. Some phototypesetters furnish standard grids with a 
variety of pi characters on the grid. Pi grids are selected for the job at hand, 
to address the unique symbols required. For example, mathematics requires 
a comprehensive set of special symbols, as does chemistry. Usually the 
superscript and subscript characters are furnished on the main grid, or ball, 
which reduces the necessity of shifting baselines. 

The typesetting of music also requires a large symbol set. 

Fonts 

Almost any typesetter can furnish adequate fonts for English language work, 
however some foreign languages strain the limits of the most sophisticated 
typesetters. In addition to requiring a very large number of symbols, often 
times the text is set in a different orientation, for example Japanese is set 
from the bottom of the page to the top, and Hebrew and Arabic are set right 
to left. Arabic is perhaps the most complex of all foreign languages to set, 
since not only are there a large number of symbols, but the characters are 
linked, similar to handwriting. 

Sizes 

In general, phototypesetters or composers can easily achieve the desired size 
by choosing the right lens. The MT/SC has difficulty with large characters, 
and in some cases (such as when printing the mathematical integral symbol) 
the MT/SC may print the upper half with one impression, and the lower half 
with another impression. Size is probably the easiest problem to overcome 
when printing complex material. 

Baselines 

Most typesetters set material on one baseline at a time, and the paper can be 


130 Composition Input 



incremented in one direction only. This means that material must be set on 
the highest baseline first, starting at the top of the page. The characters that 
are to be set appear in the center of the page or on the right, then the carriage 
of the phototypesetter must be moved over to the precise spot, the character 
set, and then the carriage returned and the paper advanced to the next 
baseline. This involves some very complex keying routines when using 
phototypesetters, although some phototypesetters offer reverse leading. To 
set a superscript symbol, the paper is simply incremented down the required 
number of points, and the symbol flashed. Then, the paper is incremented 
back to the normal baseline, and setting continues. The reverse procedure is 
followed when setting subscripts. The reverse leading feature greatly 
simplifies this multiple baseline task. 

In practice, the easiest machine to use for complex typesetting is a Selectric 
Composer or a Varityper, simply because it has visibility and an operator 
can monitor the procedure and make the necessary judgements. By using 
Varityper for example, the setting of superscript or subscript is very simple: 
when the operator reaches the point requiring the change of a baseline, she 
simply rolls the platen up or down the required distance, and sets the 
character. This is far easier than programming a phototypesetter to perform 
the precise steps necessary. Generally complex material is set by specialty 
houses, and they are equipped with all the necessary hardware to do this kind 
of typesetting on a production basis. Where smaller commercial printers 
have occasion to do this kind of work on a now and then basis, they use 
either the Selectric Composer or Varityper, or use cut and paste techniques. 


Directory of Word Processing Companies 


American Automatic Typewriter Co. 

2323 N. Pulaski Rd. 

Chicago, Ill. 60639 
(312) 384-5151 

American Automatic Typewriter Co. is a privately held corporation that has 
been manufacturing and marketing “Auto-typist” products for over 30 
years. The basic line of products is designed for the production of form or 
standard clause correspondence. Direct sales and service organizations exist 
in New York City and Chicago, with dealers representing Auto-typist in 
most sections of the United States. 

Bowne Time Sharing Inc. 

345 Hudson St. 

New York, N.Y. 10014 
(212)989-6006 

Bowne Time Sharing Inc. is a subsidiary of Bowne & Co., a large financial 
printer. “Word/One,” Bowne’s basic service, is marketed in New York, 
Boston, Washington, D.C., Philadelphia and Chicago. Through an 


Composition Input 131 



arrangement with Pacific International Computer Corp., the service is also 
available in San Francisco and Los Angeles. In addition to Word/One, BTS 
provides “Photo Comp,” a service that allows Word/One material to be 
photo-composed. 

Edit Systems Inc. 

19959 Vernier Rd. 

Harper Woods, Mich. 48225 
(313) 886-6545 

Edit Systems Inc. provides a shared logic in-house word processing system 
called “Text-Ed.” The system utilizes a PDP-8 mini-computer, which 
supports Selectric terminals and has a high speed printer. Branch offices are 
located in Detroit, Washington D.C., and New York City. 

The Edityper Corp. (a Division of Epsco Inc.) 

1335 Rockville Pike 
Rockville, Md. 20852 
(301)424-3997 

Epsco Inc. manufactures control systems, data products, and general 
electronics. The Edityper Corp. designs and manufactures word processing 
equipment. Its manufacturing facilities are located in Westwood, Mass. In 
November, 1971, Terminal Equipment Corp. announced agreement with 
Epsco to purchase Edityper Corp. Edityper’s products are marketed by 
Word Processing Products, Inc., an independent sales and service 
organization. WPPI has offices in Philadelphia, Hagerstown, Md. and 
Rockville, Md. 

Information Control Systems Inc. 

313 N. 1st St. 

Ann Arbor, Mich. 

(313) 761-1600 

ICS is in business with a product known as “AstroComp.” AstroComp 
utilizes a PDP-8 mini-computer and can support up to 8 Selectric terminals. 
A high speed printer is available with the system. Branch offices are located 
in Wilton, Conn., and Chicago to serve the New York and Chicago 
metropolitan area markets, respectively. Sales representatives also cover St. 
Louis, Boston and Washington, D.C. 

International Business Machines Corp. 

Office Products Division (OPD) Parsons Pond Drive 

Franklin Lakes, N.J. 07417 

(201)848-1900 

There are about 200 OPD sales and service offices throughout the United 
States. IBM-OPD provides extensive systems and customer support. The 
company also runs a training school (one week) for word processing center 
supervisors in Dallas, Tex. 


132 Composition Input 



Itel Information Products Corp. 

(A wholly owned subsidiary of Itel Corp.) 

2585 East Bayshore 
Palo Alto, Calif. 94303 
(415) 328-5660 

A network of about 20 branches offices covering 30 cities and about 30 
dealer organizations serving 44 cities gives Itel coverage of the major portion 
of the U.S. marketplace. Itel’s field force is second only to IBM in word 
processing. 

Lexitron Corp. 

(Formerly Autoscribe Co.) 

413 Moss 

Burbank, Calif. 91502 
(213) 843-51 11 

Lexitron builds a uniquely designed CRT display editing device. As yet, the 
company has not installed systems, but has announced orders in excess of 
$500,000. 

Metro-Tel Corp. 

409 Railroad Ave. 

Westbury, N.Y. 11590 
(516) 333-7650 

Metro-Tel manufactures telecommunications equipment and automatic 
typing equipment. Its Mate automatic typing equipment utilizes a baseplate 
that connects to most standard electrics and uses roll paper as the recording 
medium. This equipment is used primarily for form letter writing. Metro-Tel 
recently introduced Model 801, a magnetic tape cassette unit designed for 
more complete word processing applications. 

QuinData Inc. 

(A division of Quindar Electronics Inc.) 

60 Faden Rd. 

Springfield, M.J. 07081 
(201) 379-7400 

Quindar Electronics Inc. is an electronics company that manufactures 
advanced solid state equipment for use on data transmission, supervisory 
control, computer-based control and/or data acquisition systems. Quin 
Data Inc. designs, manufactures, markets and services word processing 
products. Manufacturing facilities are in Springfield, N.J. and Toronto, 
Ont. 

Redactron Corp. 

100 Parkway Drive S. 

Hauppauge, N.Y. 11787 
(516) 543-8700 

Redactron was organized in 1969 to develop and manufacture WP 
equipment. The company is presently producing both single and dual 
magnetic cassette and magnetic card WP devices. 


Composition Input 133 



Remington Rand 
(A division of Sperry Rand) 

P.O. Box 1000 
Blue Bell, Pa. 19422 
(215)646-9000 

As a leading manufacturer and marketer of office machines, Remington 
Rand has recently completed arrangements with Redactron Corp. to market 
their line of cassette and card word processing equipment. It is expected that 
Remington Rand will not substantially alter the basic Redactron product at 
the outset. 

Spiras Systems, Inc. 

332 Second Ave. 

Waltham, Mass. 02154 
(617) 891-7300 

Spiras Systems is a subsidiary of USM Corp. Spiras is a producer of mini¬ 
computers, CRT terminals and a CRT word processing unit known as 
“Accutext.” Presently, all of Spiras’ activities are located at headquarters in 
Waltham. 

Ty-Data Inc. 

109 Northeastern Blvd. 

P.O. Box 841 
Nashua, N.H. 03060 
(603) 889-1155 

Ty-Data has been manufacturing and distributing its 3600 series of magnetic 
tape cassette word processors since early 1971. The dual cassette Model 
3600/2 was introduced in October 1971. A three cassette system is expected 
by year end. Over 100 units of the one cassette model 3600/1 are in the field. 

Wang Laboratories Inc. 

836 North St. 

Tewksbury, Mass. 01876 
(617) 851-7311 

Wang is the largest United States manufacturer of electronic calculators. Its 
initial WP product entries — the Model 1210 single casette typewriter and 
Model 1220 dual cassette typewriter — combine Wang’s expertise in 
cassette, typewriter and computer technology. The 1210 and 1220 were 
introduced in late October, 1971. 


134 Composition Input 



The following four pages describe the operating 
procedures for the Redactron word processing 
system. Much of the information presented is com¬ 
mon to other systems of its type. 



!l!g§gg||glj; 





Composition Input 135 

















































































































































































GROUP 1. 

KEYS DEFINING BASIC OPERATIONS. 

Basic operations are something like those in a tape 
recorder or dictating equipment. They are indicated 
in our Group 1, at the left of the keyboard. We are 
talking about them in an order that is logical to the 
person using the machine, not the order in which you 
see them on the machine. 


They indicate the following modes: 

a. Record. Original recording of 
typewritten information. This 
takes place on the magnetic 
medium (tape or card) as well 
as on paper. 

b. Playback. Material stored on mag¬ 
netic tapes or cards is typed 
automatically on paper. "Play¬ 
back" can be stopped, and ex¬ 
tra material added to the paper, 
then started again. The extra 
material on the paper is not 
stored, however. 

c. Adjust. This is like playback, but 
automatically adjusts the right- 
hand margin while playing. 


d. Duplicate. This is operative on!, 
in two-cassette or two-card edg¬ 
ing typewriters. Information or 
one cassette or card is trans¬ 
ferred to the other at very high 
speed, without printing on paper. 

e. Edit. This is like playback, but 
when printing is stopped, the 
operator can type in new ma¬ 
terial, which is then part of the 
magnetic record as well as ap¬ 
pearing on the paper. 

f. Skip/Delete. This, too, is a kinc 
of playback. "Skip" is used with 
"play" or "adjust" to bypass 
material on the tape or card — 
that is, not print it. But the 
skipped material stays on the 
tape or card. When "skip" is 
used with "edit", however, ("ed¬ 
iting" being another word for 
correcting), the material is ac¬ 
tually deleted or removed from 
the tape or card. 


I RECORD| 


BACKUP 



WORO END CONT 
UNDER UNDER UNDER 


REOD MARK 



SPACE 


GROUP 2. 

ACTION KEYS. 

They start the tape or card in motion and define 
where the action stops. You can think of them also 
as positioning keys. 

—When the basic operation is "record", the action 
keys in the row under the reverse light cause the tape 
or card to back up to the specified position. 

—When the basic operation is "playback", "adjust", 
"edit" or "skip", the tape or card will go forward. 

—When "edit" is the basic operation, and the "record" 
button is depressed, the keyboard will lock and the 
operator may go back by depressing a suitable 
action button. 


a. Character/stop. The machine 
stops at the next character. For¬ 
ward or reverse. 

b. Word. The machine stops after 
the next detected space forward. 
When backing up for correction 
it stops at the first letter of the 
word just passed. 

c. Line. The machine stops at the 
next carrier return in forward. 
When backing up, it reverses 
one physical block. 


d. Find Mark. The machine stops 
at the designated "mark". In the 
tape machine, the "mark" is the 
typist's notation to herself about 
a place where she wants a stop 
Marks may be "counted" b 
the machine so that they can be 
returned to by number, at ver, 
high speed. Forward or reverse. 

In the card machine, "mark 
sends you back to the begin¬ 
ning of the card. 

e. Para(graph). The machine (go¬ 
ing forward only) stops at the 
beginning of the next paragraph, 

f. Auto. The machine types for¬ 
ward and stops at the next 
"stop" character. 



















GROUP 3. 

KEYS FOR SPECIFIC INSTRUCTIONS 
TO THE SYSTEM (CONTROLS). 

The keys that are control keys are all regular type¬ 
writer keys as well. To function as specific instruc¬ 
tions to the machine, the "Code" key has to be 
pushed along with the regular typewriter key. Most of 
these keys' additional jobs are indicated on a horizon¬ 
tal bezel above the keyboard. 

Control characters are recorded as single special 
characters on the tape or card. If it is desired that they 
print, as well, the print control switch (labeled PC) at 
lower left is switched on. The character then prints 
with a slash on it like this: 2 (The slash of course, 
indicates that this is an instruction code, not the 
ordinary typing function.) 

The control keys are divided into two sets: the keys 
we know as format keys (that is,ones we ordinarily 
use, like spaces, backspaces, carrier returns); and the 
keys we know as number and symbol keys — the top 
line of the typewriter. 


Taking the formal format keys 

first, and from the bottom up: 

a. Required space (/). This is a 
space that must be considered a 
graphic character. It inserts spa¬ 
ces between words that must be 
on the same line, as, for example, 
in a date, June 22, 1971. 

b. Required carrier return (RCR). 
This means that the carrier re¬ 
turn signal can't be eliminated 
in "adjust". This might be used 
for the three or four-line ad¬ 
dress on the top of a letter. 

c. Required special carrier return 
(SCR). Like the required carrier 
return, but the line is not re¬ 
corded. Used during 1-card, edit, 
where it is desirable to get a lot 
of information on the card, but 
where the width of the line will 
be narrower on the paper. 

d. Required backspace —). This is 
needed to make characters not 
on the standard keyboard, as 
for example, the accents over e 
in some languages. 

e. Automatic tab. The typewriter 
will automatically execute that 
tab after every carrier return. 
"Required carrier return" dis¬ 
continues automatic tab. 

Now, taking the character keys 

that normally serve graphic func¬ 
tions: 

a. The H is recorded as (£) Mark. 
This is used for separating blocks 
of material. 


c. The (0) is for starting continuous 
underlining in groups of words 

d. The (9; is for "end underline", 
meaning stop the underline. 

e. The ($) is for starting word by 
word underlining. 

f. The (2) is a Stop code, stopping 
the machine during playback and 
edit when Auto action is used. 

g. The (0) is a character called Link 
that defines the end of a block 
without generating a Carrier Re¬ 
turn. If it is used with the Special 
Carrier Return, it permits cram¬ 
ming the magnetic medium with 
data that will print in the desig¬ 
nated format. 

h. The (8) is an "end of tape" (EOT) 
code, for the tape machine. On 
the card machine, it means eject 
card. 

i. The ( 4 ) activates switching from 
one card or cassette to the other. 
This, of course, is useful only 
when there are two. The charac¬ 
ter may also be called Alternate. 
It can be operated in "play" or 
"adjust" if you want to manually 
switch from one card or cassette 
to the other. 

j. The (3) is a repeat signal. It will 
repeat the Mark previously read 
or the entire card. 

k. The (2) is for double spacing 
material. 

l. The (^) is for single spacing 
material. 


b. The (-) becomes a required hy¬ 
phen (/) when coded. A required 
hyphen is part of a word that 
cannot be eliminated if it is run 
together (unlike a regular hy¬ 
phen at the end of a line, but in 
the middle of a word). Example: 
son-in-law. 



































































































































GROUP 4. 

ADDITIONAL KEYBOARD INDICATORS. 


a. At the lower left: the print 
control, determines whether the 
codes are printed, as well as 
recorded. 

b. At the upper right: the two keys 
designated Units and Tens are 
the Mark designating buttons for 
the Autofind. These are used to 
set the work numbers for the 
"Find" counter on the console. 
To use them you must depress 
the code key. (Tape only.) 


c. The light at the right above the 
keys is the Back-up light. When 
that light is on, back-up direction 
is indicated. The buttons immed¬ 
iately below it cause the tape or 
card to go back to the positions 
described. 

d. The lefthand light on the key¬ 
board is the Record light, mean¬ 
ing the tape or card is going 
forward. When the light is on, 
the information keyboarded is 
being recorded. 



CONSOLE INDICATORS: 



Console — 2 cassette model 


ADDITIONAL INDICATORS 
ON THE CONSOLE. 


The Redactron editing typewriter (single and double card; 
single and double tape versions) has a minimum number of 
controls on the console. 

The card console has an indicator (for track number), and 
three buttons: card eject, track up and track down. 

The tape version has buttons for: Rewind and Eject, 
and the indicator for Autofind. This is an option. 


2 6 7 
□ □□ 
4 


2 6 7 

□ □□ 

4 


Console — 2 card model 


1 Rewind 

2 Eject 


5 Track Indicators 

6 Down 


3 Tape Gates 7 Up 


4 Card Gates 8 "Find" Counter 

Display (2 Digits) 











































































































































































10. Keyboard arrangements 


The following pages illustrate some of the keyboard layouts that are 
available for composition input. Detailed information on layouts and 
operation for specific applications may be obtained directly from the 
manufacturers. 


Composition Input 139 




140 Composition Input 


This is the basic layout designed specifically to drive 
a Linotype. 

























JLJ H— LJU 
- O' —■' 

Z. lu 3 

U > 0£L 


Q , Qh 

DiuDO 

O-'OE 



Q- Q 

O < 


f— LU 

co q: 


a: 3 

< < 

|_ UJ 

CO « 


Q a. ■ 

-l CO 

Z Z 

LU LU 

< <3 ’ 

* > 

> o 

VP ® ' 

h 

t£ 0> 

af 

1 oo 

* <9 
•i N 

* a. 

% VP 

dJ 

tt in 

o o 

* <r 

o 

cv.cn 

5" co 
$ °* 

THIN 

SPACE 


CAR 

RET 


CL 

— - <n> 

CL 

O 


D 

> 

► 

h- 

Ql 

UJ 


* 

Cf 


z y= 

=5 X 
CO 


X 

CO 


H ® 

in 

« <l • ^ 





I 

1 

U. 

O 

in 

< 







2 


<0 


> 


u 


X 


IM 



H— 

LL. 


X 

h— 

CO 

z ^ 


3 X 


CO 



fc 2 

> 3 

8 - 
l-H 

§ u 

2 U 

Ph 

<L> 

<D J 3 
JG *5 

4-* 

a_, •» 

o 1/3 
O 

3 **■* 
O 

Cd <D 

»—I 4-* 

T3 2 

Vh 

cd <u 
O X 
X> ^ 
>* 

<D cj 
44 — 


<d 

43 

4-» 

(A 
• 1 "* 

CD 

• *■* 

43 

H 


T3 

<D 

O 

G 

TD 

O 

Vh 


<D ' 
Vh ^ 

Cd g 

a, > 

£ fc 
a> 

2 « 
43 
>» Uh 

»—H L 

’S C 
o 5 

5r< O 


O o 

Vh <m 

G to 
O W 

u & 

Vh ~ 

2 g 

S3 g- 

u .S 


£ CD 

13 2 

2 cm 
2 O 

CD M 

43 cd 
H 43 

^ 4-* 

C/) Q 
<D 4-* 

O 

,cd ^ 

cm or) 
<D O 

£o 

Z £* 

G <D 

V > 

.13 or) 
53 5 
? * 

O Vh 
G cd 

si 

CD 

T3 44 
<D 

to u 

3 u 

S3 J 

£ CD 

m 43 
G 4-» 

O ^ 

•<“H Tr 

</) O 


Composition Input 141 


Linotype. This particular layout includes small caps. 
There are still quite a few LCCs around, and at used 
equipment prices, they can do a satisfactory input 




PUNCH TAPE I CODE START STOP STOP LINE I J-CAR 

ON FEED I DEL REAO READ CODE DEL I RET 




142 Composition Input 


The layout utilized a standard typewriter layout with 
distinct shift (upper case) and unshift (lower case) 
keys. Every line had to end with a “J-Car Return” as 
tapes were being prepared. 






Composition Input 143 












































































WORD 

DEL 

(OPT) 

QUAD 

LEFT 

QUAD 

CENTER 

ADD 

THIN 

RETURN 

OR 

RETURN 

PLUS 

ELEVATE 


x o 
O < 

c* cm ^ 

Q O “ 

c LU 
-S CL 
2 CL 
O => 



<2; 






«£in 






N^CVI 


> 


H 




UJ 


£ 


o 



uj £f 

Q uj 

O — , 
n lu 

u Q 


> UJ Q 

LU UJ 

a < JP 




a 

tv.TSl- _J 

- o- 

c#o)- y 


I <x> 



X 


0 


IL 


Q 


If) 


< 


H- ^ 

± u 
X O 

co —i 


m q 

CL LU 
< LU 


a 

D oo 
< 


a 

Q N“ 

< 00 


2 


Z 



> 


u 


X 


N 


<* . 
lu zd 
CL. < 


DC QC ~ 

UJ o £ 


Q ^ 

C£ — J H— 

O LU CL. 

^QO 


LJ LU 


x o 
0 < 
Cm ^ 

Q O oc 

< LJ uj 

-S o- 
0 3 



^ Q- 
m CO 


Z CL 
LU CO 







O 

< 

£ 

&: cm 

w O 
O 


^ Q - 
m co 


Z CL 

LU CO 


<0 00 


c* i-* 


j vO 


3 


> 


h 


Z 


X 


0 



•n lO 




m CO 



a 


LU 





> 


o 


a 


co 


X 


N 


? o. 
X co 



LU “ 
Q LU 

O — 1 
u Q 


> UJ Q 
LU UJ 

rv' < UJ 


2^ 
X CO 


cm 

LU _J 

^ < 
O 04 



Ck CQ 

Q 

< ~ 


Q CQ 

Q 

< hs 


s- < 


^ < 
O £ 


PAPER FEED 



















































































































































H— O 
Q Z 
Z < 
— X 


< i 

u O 


fV LL 

o ^ 


O Q 

a: < 

LU LU 
M —i 


o 9 

Q < 
< ^ 


z ^ 

— M 

O CO 
CL 


LJJ UJ 

Q- U 

>- < 


II 

88 


^ x 

U UJ 
LU h- 

cm 


O 





I 00 




;£<o 


&\n 









5 




04 


3 


> 


h 


a: 


UJ 


£ 


X 


o 


Ll 


D 



CO 


< 


> 


O 


X 


N 


* 


LU 

5 : U 
< < 
W Q_ 

CO 


LU 

Z U 

^ 2 

CO 



UJ Q 
CL ill 
< LU 



g £ 

O ^ 

> m 
> Q 


CO h- 
< LU 
H— CO 



I— < 

O cm: 

Z ^ 

— CL 


Z P 


Q u_ 

Z “Li 


- O 

9 z 
± < 

X 


a ^ 

Z £ 

LU O 


< I 


z “ 

I I I 


0 9 

cm < 

UJ LU 
M —' 


a Q 
a < 
< “J 


z ^ 

^ UJ 

LL. Q- 

S o 


z m 

_ M 

O CO 
Q_ 


UJ LU 
CL U 
>■ < 


CO ^ 

o o 

U u 


uj hH 
-7 > 
± < 
- i co 


£ < LU 
“■ h- u_ 


UJ £ 
Q m 

8 =J 

u Q 



< £ 

3 uj 

0 - 1 


Q S 

^ < 
lu l_: 


_ oc 

Q LU 

< t- 

3 Z 

0 m 
u U 



O t 

^ "E 

3 O 

0 DC 


o =: 

< Q 

3 Q 

oi 




o £ 

£ O 



—w 


C'"’G 



<* 0 > 


I 00 






u^cO 











£ 


N 


< 



LU 

^ u 
< < 
LU CL 

co 


LU 

z < 

^ 2 

CO 


UJ Q 
CL m 
< LU 
\— LL 


rv CO 
< ^ 
5 •" 


CO < 

< 2 
*- 0 
LU 


9 ^ 

< LU 
I— QC 



£i £ 

3 X 

co <0 


c* (— 


3 “ 

< z 

»— LU 

V 


co X 

< o 


5 Q 

< Q 


CO £ 
lu 


3 5 

< UJ 

1 — cm 



































































































































=! Z 

^ “3 


CXL 

< •“ 
CJ lli 
V on 


73 

« ■ 

C 

43 g 

03 

~ o 

/—V 

«4-< ° 

CD 

o 

G 

00 

o 

«3 C 


G *r7 

<N 

o .s 

O 

«o £ 

(N 

i-i *•* 

CD 

00 

> «-t 

u 

£ 
»-i ^ 

+■» 

CD 

•N 

eS ^ 

C/3 

43 CD 

'O 

73 

Ui 

cd 

fli cd 
u u 

O 

03 W> 

JO 

Ou 

<D 

CD 

43 no 

a 

H £ 

<L> 

cd 

73 

. O 

• 1-4 

C" 43 

u 

o >> 

Ph 

£ (D 

o 

H M 

£ 

s—✓ 

Ui 

<D 

V-t 

cn a 

S ’C 

cd 

<N > 
00 > 

<D 

o 

U. 

4-* 

<D 

<D a 

as 

43 3 
+-» ►—> 


o 

a 


o 

£ 

H 












































































a % s 

U - £ 



specific unit. Note that the “Insert Leader” and other 
functions are single keystrokes. This keyboard is 
made for Compugraphic by Automix. 















































z 

o 


OH LU 

o z 

X X 

co 


O OJ 


£ o 

i— cm 
Z < 

“ £ 


U _ tz UJ 

Eu Z Z x 

O O a: 5: 

£ 



“ £ 
< t— 

x u 
u < 


i o 

Cl 



lu 

z y 

LU Z 
CO LU 



«$N 


- iO 


£tO 


«o-sf 


•ft**) 


D 


> 


H 


q: 


5 cj 


LaJ 


£ 


a 


x 


0 


Ll 


a 


</> 


< 


2 


CO 


> 


a 


x 


N 


z u 

LU Z 
CO LU 



Cm 

•m CT 

t; ^ 

• »H 

■M G 
CJ 

CJ 

CL 


CJ o 

•M 

M O 

Vh +-> 

Cm 

O 

/■N #f,H 

<2 s 


TD Oh 
<L> CJ G 

M tt) ** 

goo 

° o 
* ■■ ^ 

«»*2 g 

CO t« £J 
^ c« ^ 

^ ^ -g 

<D G 
tj ~ CO 

i-j c« 

£ u “> 

O -*-» G 

x *c *-0 
JJ £ - 

Uh OJ -O 

a „ 

bo 

G +-* rj 

CJ f\ • 

*T-t -M 

g 'X -r 1 

Lh CO g 

e > 

„2 5 


LU Q 
Q- LU 
< LU 


Q- ^ 

28 

co (j 


laj a. 

o_ — 

< * 
m co 


£- Q 

O < 

h- LU 

co cm 


E 2 
< < 
h- ULJ 






















































•- 1 X 

. « 


u 

o 

CO 

CO 

<D 

O 

o 

u 

Oh 

T3 

Vh 

O 

£ 

W 

H 

HH 

<D 

43 


CO 

T3 

T3 

cd 

T3 

G 

cd 

T3 

CO 

cd 

G 

O 

O 


• 

W 

5>> 

O 

<D 

c 

44 


O 



-«-» 

O 

o 

<D 

S-H 

•»-* 

a 

13 

o 

C/5 

o 


3 03 

o HH 

*t3 
cd £ 

cd 

« 12 
■S S 


co 


co 
cd 

.2 «> 

_S « 

p 5 CO 

H ss 



Composition Input 151 








































Expanded 



cm 

LU 

£ 

o 

Q- 


2 r => 
< < 


cm 

LU zd 

Q- < 


X 

lo 


T3 

Lh 

cd 

T3 

C 

cd 

-♦-* 

CO 


cd ^ 
~ cd 

.52 a 

cd 

O 

O (30 
<N G 
o° -g 

e *p 

<i> c 
O vm 

■C ° 

<L) 

<\ Lh 

2 w> 

fC a} 

H *5 


c 
o 

• M 
4-* 

• 

o 
a* 

S 
o 
o 

X 
<L> 

a. . 

S G 
o .2 

O CA 


CD •— 

3 3 
O' O 


• • 


o 


•« <T> 


1 CO 




^ vo 


S in 


S. 4* 


CO 


<M 


<2 



cc 

Z Q 

UJ —i 

+ 1 

a 

o 

— 

D 

> 

1- 

a: 


UJ 


£ 

O' 

EN 

SPACE 

THIN 

SPACE 

o 

a 



Q_ 

o 

1— 

00 

s 9 

LU —J 

H— 

LU 

cm 

a 

-i 

sc 

-> 

X 

o 

LL 

a 

vO 

< 

EM 

SPACE 

VERT 

RULE 

oc 

a 



u. 

a. 

TAPE 

FEED 

> 

LU 

—J 

LU 

#• • 

UN 

SHIFT 

2 

Z 

CD 

> 

o 

X 

M 

h— 

LL. 

X 

to 


CC 

D 


Q Z 
< h— 























































































justified cards 


The Random Access Composition Entry system which uses punched cards as input to a 
phototypesetter. 



Memory capacity can be increased by adding discs or tape units. 


An editing system. Non-justified input is “massaged” by keyboard and video terminal and 
output as justified or unjustified tape. 


Wire Service 
Lines 



Another editing system. Wire Service input is received by a computer and allocated to one 
or several video terminals by a controller. After editing is completed, data is output via 
punch. 


Composition input devices become part of systems. 


Composition Input 153 

















A formatting system. Unjustified data is read into a computer with the capability to 
hyphenate using a dictionary and to access a large library of type font widths. Justified 
and formatted tapes are then produced to run a particular phototypesetter. 




non-justified 
mag tape 


(or) 


justified tape 
mag or paper 


phototypesetter 

accepts non-justified 
mag tape 


Linecaster or 
phototypesetter 


A magaetic tape input system. 



justified tape mag 
or paper 


\ 



Phototypesetter 
accepts non- 
justified mag tape 


A magnetic tape cassette input system. 


154 


Composition Input 





















COMPANY 


Alphatype Corporation 
7500 McCormick Blvd 
Skokie, III 60076 
Phone (312) 675-7210 


Automix Keyboards, Inc 
13256 Northrup Way 
Bellevue, Wash 98005 
Phone (206) 747-6960 


Composition Systems, Inc 
325 Central Ave 
White Plains, N Y 10606 
Phone (914) 761-7800 


Compugraphic Corp 
80 Industrial Way 
Wilmington, Mass 01887 
Phone (617) 944-6555 


CompuScan, Inc 
900 Huyler St 
Teterboro, N J 07608 
Phone (201) 288-6000 


Computer Composition, Inc 
16661 Ventura Blvd 
Encino, Cal 91316 
Phone (213) 986-5552 


Datatype Corp 
1050 N. W. 163rd Dr 
Miami, Fla 33169 
Phone (305) 625-8451 


Digital Equipment Corp 
146 Main St 
Maynard, Mass 01754 
Phone (617) 897-51 11 


ECRM, Inc 
1 7 Tudor St 
Cambridge, Mass 02139 
Phone (617) 661 8600 


ETTA Associates 
22 Millbrook Lane 
Wakefield, Mass 01880 
Phone (617) 246-1384 


Graphic Products Corp 
522 Cottage Grove Pd 
Bloomfield, Conn 06002 
Phone (203) 243-0730 


Graphic Systems, Inc 
217 Jackson St 
Lowell, Mass 01852 
Phone (617) 459-2111 


Harris-lntertype Corp 
55 Public Sq 
Cleveland, Oh 44114 
Phone (216) 861-7906 


Hendrix Electronics, Inc 
Grenier Industrial Village 
Londonderry, N H 03053 
Phone (603) 669-9050 


Imlac Corp 
N E Industrial Center 
Needham, Mass 02194 
Phone (617) 891-1600 


Information Control Systems, Inc 
313 N. First St 
Ann Arbor, Mich 48106 
Phone (313) 531-5032 


Interface Mechanisms, Inc 
5503 232nd St, S. W. 

Mountlake Terrace, Wash 98043 
Phone (206) 774-3511 


Mergenthaler Linotype Co 
Mergenthaler Dr 
Plainview, N Y 11803 
Phone (516) 694-1300 


MGD Graphic Systems 
3100 S. Central Ave 
Chicago, III 60650 
Phone (312) 242-4860 


Omni-Text, Inc 
406 W. Washington 
Ann Arbor, Mich 48103 
Phone (313) 769-4826 


Photon, Inc 
355 Middlesex Ave 
Wilmington, Mass 01887 
Phone (617) 933-7000 


Redactron Corp 
100 Parkway Dr, S. 
Hauppage, N Y 11787 
Phone (516) 543-8700 


Singer Graphic Systems 
2350 Washington Ave 
San Leandro, Calif 94557 
Phone (415) 357-6800 


Star Graphic Systems 
2 S. Main St 
Hackensack, N J 07606 
Phone (201) 652-7200 


Tal-Star Computer Systems, 
10 Lake Dr 

Hightstown, N J 08520 
Phone (609) 799-1 1 1 1 


Varisystems Corp 
80 Skyline Dr 
Plainvievv, N Y 11803 
Phone (516) 931-7200 


VariTyper Division, A-M 
11 Mt Pleasant Ave 
Hanover, N J 07936 
Phone (201) 887-8000 


Warlock Computer Corp 
Route 7 

Georgetown, Conn 06829 
Phone (203) 544-8308 


Xylogic Systems, Inc 
13 Mercer Rd 
Natick, Mass 01760 
Phone (617) 655-5800 


(made by 
Graphic Systems) 


















































■M 

CD 

E 

03 


C 

o 

CD 

H— 

CL 

O) 


C 

CD 


C 

*03 

i— 

C 

03 

CD 

X 

UJ 

x: 

o 


o 


w w in i/i vi 

CD CD 03 03 CD 

> > > > > 


n <✓> </> 

p fl) QJ 

C > > 


000^5^0000 

ccc^^cccc 


o o o 
c c c 


t/3 t/3 t/3 t/3 

03 03 03 03 

> > > > 


$ s g ° g 
> > c c 


o o o 
O if) ld 
o_ o CO 

co oo" cm" 

CM 


IT) o 
px , o 
co , o 


in in in tn 
0 ) 0)0 0 ) 
^ 

cm" cm" cm" cm" 


o o o o o 

o o © o o 

o_ o_ o_ o_ o_ 

rx' rx' rx" tx rx' 

ld id ld id if) 

to co co co cd 

CO CO CO CO CO 

cm" cm" cm" cm" cm" 


■|?3 

2-^0 

! x ° 
O X o 

n. h- CD 


>>>>>> c c 

CCCCCC 0503 

<<<<<<>> 


a 

a 

03 

■*-< 



E 

E 

1— 



o 

o 




o 

o 

o 

CM 


i_ 

k- 

■4—* 
t/3 

O 1 

o 

CD 

CD 

D 

CN 1 

CN 

> 

> 

“D 

CO 

CO 


0)0 0 ) 0 ) 
CL Cl Q. Q. 

O O o o 

4-J 4-» 4-» 4-> 

D D D D 

< < < < 


03 0) 0) 0) 0) 

1/3 t/3 t/3 t/3 t/3 

o o o o o 

l. k- i_ L- u. 

+-> 4-< +-* +-< 

o o o o o 

03 03 0) 0) 0) 

UJ 11! UJ UJ UJ 


o o o o o 
o o o o o 


o lo o 
CD CM O) 


oooooooo 
oooooococo 
co co if) if) if) o 


o o o 

i—. r—■ 


o o o o 
o o o o 

CM CM CM !N 


O O O O O 

If) If) If) If) ID 


O O O ID O O 

ID O t- CM O 

«- T- CO 


o o o o 

LO ID O O 

t- ID CD 


O LD O 
»- CM LO 


O O O LO LD 

ld o id rx rx 


o o lo o 

LO LO fx LO 


0 , 0 , 

O , id , 


03 

X) 

E 

■a 

03 

a 

o 

LD 

LD 

o 

o 

o 

O 

o 

O 

o 

o 

D 

a 

o 

CN 


o 

LD 

o 

O 

LD 

O 

o 

LD 

X 

X 

CO 

to 

CM 

LD 

r— 


CN 

o 

*sf 



O) 



O O ID 
O L0 CM 
o- CO 


OlDOOOOOOO 
IXLOCMOOlOCMOLD 
«- >- CO CO 00 


o o o o o o 

o o o o o o 

CO CO o- CO 


o o o o 

O O LO LD 


o o o o 

O O O LD 
CO CM o- «- 


o 

o 

o 

o 

o 

o 

o 

o 



CO 

CN 

CM 

CO 

O 

<— 

CO 

o 


O 

O) 

CO 

LD 

f^. 

r>x 

r^- 



r>. 

p*. 

r^. 

r^. 


r-x 


r-x 

CD 

PV 


CD 

r*x 

LD 

P^ 

CD 

CD 

CD 

OJ 

a> 

0 ) 

® 

a) 

® 

03 

O) 

03 

03 

0 ) 

® 

® 

0 ) 

O) 

® 

0 ) 

03 

® 

O) 

O) 

03 

O 


®ooo*-»- 

co rx rx ix rx rx 

O) 0 ) O O O) o 


O O O r- 
rx ix rx rx 
0 0)0)0) 


Q O ^ r- CM 

cd rx rx rx rx 

o o o o © 


OOOOOOOOOOO 
LOLOLOOLOLOLOLOlOlOLO 
0) C) 0) D G) O) G) O) O) C> 
|x" |x rx O’" O O’ O’ CN O CD «— 


O O O 

o o o 

ODD 

D O’ <— 
CD D i- 


O O O O O O O -vT fo - 
OOODOOO^SS 
o-oocmooo c ? ic ) j 
d" cm" cd" <-' d" d" 2 2 

CDO-DCMO-CDCD22 


o o 
o o 

D D 

cm" cd" 9 
™ ™ $ 
° ° co" 


o o o o o o 

D D D O O O 

O O O D D D 


o o o o 
o o o o 

D O CM D 


D CD CD O O’ CD CM N O’ CM 

CM CM CM i- <- r- CM 


O O O D O O 

O D D CO D O 

LD_ 0) O) 00 0) LO_ 

co" rj" r-'" jx." 

CM «- r- CM «- 


t/3 t/3 t/3 O t/3 

03 03 03 r- 03 

> > > > 


o o o 
c c c 


o <33 
C 03 03 


1/3 t/3 CO C/3 1/3 t/3 O 

03 03 03 03 0 03 r- 

>>>>>> 


t/3 t/3 t/3 

03 03 03 

> > > 


t/3 t/3 t/3 t/3 

03 03 03 03 

> > > > 


t/3 t/3 t/3 t/3 t/3 

03 03 03 03 03 

> > > > > 


cd cd cd 


"0 

“0 


•0 

■0 

■0 

k_ 

u. 


k. 

k. 


CO 

CD 


CD 

CD 

CD 

O 

o 

-Q 

o 

o 

o 

X) 

n 

o 

-Q 

X) 

X) 

> 

> 

C 

> 

> 

> 

03 

03 


03 

03 

03 







+-» 

4-J 

4^ 

4—* 

4-> 

4— 

o 

o 

a 

o 

o 

o 

03 

0) 

03 

03 

03 

03 


u. 


k. 

v- 

k. 

Q 

b 

b 

b 

5 

b 


CD CD CO CO CO 
q * * ' * CO - 

_q co co r^. 

S' cd" cd" cd id id . id 

CD 


|x CD CD CD CD 


CO CD CD CO 


co oo co oo oo _5 

i*'* rx' rx" rx' rx" > 


tO CD CO CD CD 


S I 8 

J 0) Q. 


CO LD LO CO LD 
CO -*fr CO 


LD LD CD 
^ 


^ CM CN 

ld 


tf^COLDLDLDLD^'sf 
LOLD^^-'sJ'^^LDlD 
LD LO 


LD LD LD 
^ ^ 


CO CO CO LD 
CO CO CO ^ 


LD CO CO LD LD 

^ co co 


CN^^tCN^CNCNCNCN^ 

^(NCMr-cNrNr-rsrsCN V 

LDLDLDLDLD^-O'Sj'CDCD 
r- CO t- 


CN CN 

cn 


CNCNCNCDCDCDCDCNCN^- 

r^r^r^cocococor^r^^ 

CDCDCDCDCDCDCOCDCDCO 


CM CN CD CD CD CD 
r- r- C[) CO CO 

LD LO CN LD LD LD 


^ co 

CM CN CM 
LD LD ID ID 
LD LD LD 


CO CO CO CO CM ^ 

T-T-T-T-r^T- 

CD CD ID LD LD CD 


r-r-r-r-CNCOCOCOCN 


CMCM^COCOCOCO^^CO 
T- T- CN T T ^ 


r- *— Q O LD LD 


O) CN ^ O CO 
CM 


OOOOOCMCNCNCNOO 

0)0^0)0)0)00000)03 


O O O 
CN CM CM 


OOOCDCDCDCDCNCMO 
0)0)0)0)030)0)'— *— O 


O ^ CN CM CN 
03 + 0 0 0 


O O O 
0)0*- 


CN O CN CN CN O 
*- O O 


CNCNCNCN^^^^COCM 


LO 

LD T CD 
*- O 


CDCDCN'tOOCO’tfCOOO^ 

^ <3* 


CN CN ^ ^ ^ 


CN CM CN CD 


CO CN ^ CO ^ 


O LD LD O LD 
CD CN CM CO CN 


O 

O LD o 
CO *7 LD 
6 
o 


o 

ldldooooo|£ 

CN CN tt O O LO ', 
"—CM O 
® 


If) If) Lf) 

CO 00 CO 


o o o o 

if) rf 


o o o o o 
cm <a- in oo if) 


.% 2 g 8 I 

CO —I —I —I'D .E o § 

XH-H b- X q O) q. 

t— t— t— 0*”00 -m^ c 

cDCDr^r^cDOO of; £ 
OOOOJOCMCMI^O 
CMCMCNCN^fr^r^Q-<lCJ 


o u- 
° C o 

CN >< O 
r- h (O 


c c c 
o o o 


o o o 

UL LL LL 


o o 
o o o 

LT) r- r- CN 
CN CO CO CO CO 


O'— CN CO CO CO CO u o % 

O CD CO '— '— '— '— CD CD ^ 

CNLDLor^r^r^r^o-Q-X 


•D o O O 
^ o o o 

> O *- CM 

ro oo co a) 

CL x X X 


X X 'P 'P 'P 

03 CD “o c C C 

^ ^ o o o o 

ts « O o o o 

0 d r r r r 

-> -3 CL Q_ Q_ CL 


D _J 

x x Q 

O O O r- 

ro ® ® ® 


in rx n- rx co c 

CMO'f'^M’i- 
cx rx. rx. ^ 

2^225 S 

< < < < < X 


156 Composition Input