In researching the early days of word process-
ing software for personal computers, I have
examined the relationship between software
capabilities and microcomputer hardware im-
provements, including the impact of operating
systems, storage devices, and printer technolo-
gies. Moreover, in this article I document the
growth of the microcomputer marketplace
and the role of trade publications and user
My primary focus here is on the ear-
liest software for PCs: Electric Pencil, Easy-
Writer, and WordStar.
A companion article in this issue (“The
Proliferation and Consolidation of Word
Processing Software: 1985−1995”) examines the
successors to WordStar in the IBM PC environ-
ment: WordPerfect and Microsoft Word.
The evolution of microcomputer word pro-
cessing software is an example of a serial monop-
A monopoly occurs when one company
dominates a market segment over a long period.
In a serial monopoly, companies succeed one
another as the dominant player in the market-
place. WordStar became the dominant player
in the CP/M (Control Program for Micro-
computers) environment. In the DOS environ-
ment, WordPerfect succeeded WordStar. When
Microsoft Windows supplanted the DOS envi-
ronment, Microsoft Word for Windows took
over as the leading word processing software,
which it remains today.
The history of microcomputer word pro-
cessing software can be divided into three over-
lapping phases: origins, proliferation, and
consolidation. The origins period began with
the Altair 8080 computer and the first word
processing software for microcomputers,
Electric Pencil. During this period, microcom-
puting per se was dominated by hobbyists
supported by groups devoted to a specific
computer, such as the Altair, the Sol,
Apple II, and the Radio Shack TRS-80. Once
the CP/M operating system became the de
facto standard, these individual groups of users
became part of broader communities, and the
software market moved away from vendor-spe-
cific environments to increasingly standard-
ized versions of CP/M. This period was also
marked by the new technical periodicals and
the growth of a retail industry.
The proliferation phase began with the 1981
introduction of the IBM personal computer.
During this phase, the IBM PC (PC/MS-DOS)
standard became a signiﬁcant market presence.
This period saw fierce competition between
vendors of increasingly powerful word process-
ing packages. The consolidation phase started
with the introduction of Microsoft Windows
3.0, in 1989. This version of Windows received
“unequivocal market acceptance,” and accord-
ing to historian Martin Campbell-Kelly, “a par-
adigm-shift … occurred.”
My focus here is on
the origins phase; the follow-on article in this
issue will delve into the proliferation and con-
Throughout both articles, we will watch a
rapid evolution—from the limited microcom-
32 IEEE Annals of the History of Computing Published by the IEEE Computer Society 1058-6180/06/$20.00 © 2006 IEEE
The Origins of Word Processing
Software for Personal Computers:
Thomas J. (Tim) Bergin
Since Electric Pencil ﬁrst debuted in 1976, more than 400 other word
processing packages have emerged, most fading into oblivion. This
article—ﬁrst of two in this issue that together recount the history of
microcomputer word processing software—focuses on three of the
earliest word processing software packages, Electric Pencil,
EasyWriter, and WordStar, which was the mid-1980s leader in the
CP/M, PC-DOS, and MS-DOS operating system environments.
Visit the Annals Web Extras archive at http://www.computer.
org/portal/pages/annals/content/webextras.html to access the
“Word Processing Timeline” that accompanies this article.
puters used by hobbyists in the 1970s, to the
increasingly powerful personal computers of
the 1980s, culminating in the ubiquitous infor-
mation appliances of the 1990s. Although the
improvements in computer hardware have
been impressive, the evolution and prolifera-
tion of software, especially word processing sys-
tems, is nothing short of phenomenal.
In the beginning
In the 1970s, large organizations depended
on mainframe and minicomputer-based systems
to accomplish their computing. Consequently,
the appearance of the Altair 8800 on the January
1975 cover of Popular Electronics, an electronics
hobbyist magazine, was hardly noticed by these
organizations—or by the systems analysts, pro-
grammers, and machine operators who worked
in their computer centers. The Altair 8800
would start the microcomputer revolution and
spawn an ever growing hobbyist community of
Moreover, the Altair’s
success encouraged creation of a number of
start-up companies to build microcomputers.
Among these were IMSAI Manufacturing,
Cromemco, Processor Technology, Apple
Computer, Commodore, Radio Shack, and
Southwest Technical Products.
To make these new computers useful, cre-
ative people wrote software for the new micro-
computer hobbyist community. The early
periodicals were stuffed with ads for games;
household applications like recipe manage-
ment; and business applications such as spread-
sheets, word processing, and ﬁle management
programs. As Campbell-Kelly observed:
The personal computer word processor originat-
ed not as a replacement for the dedicated word
processor but as a useful application for users of
home and hobby computers. The ﬁrst packages,
such as Electric Pencil and EasyWriter, were
developed by lone authors in 1976 and 1977.
The creation and maintenance of word pro-
cessing programs exploded into a subindustry.
Within ﬁve years, 100 or more word processing
packages were being offered, and by 1984,
there were 300 word processing packages for
the IBM PC alone.
Necessity as the mother of invention:
Michael Shrayer (see Figure 1) was a pro-
grammer who attended meetings of the
Homebrew Computer Club.
At a fall 1975
meeting, an attendee, Bob Marsh, offered
everyone a copy of a software package that his
company, Processor Technology, had devel-
oped, aptly named Software Package One
Processor Technology had contracted
with Jerry Kirk and Paul Greenfield of
MicroTech to create a set of programmer’s tools
to make it easier to write, edit, and debug pro-
grams for the Sol computer. Marsh’s partner
Gary Ingram improved the toolset and named
it Software Package 1.
Because Shrayer was in the middle of docu-
menting a programming project, he thought it
would be a good idea to add code to SP1 so that
he could use it to help him create and edit the
project’s documentation. The result was
Extended Software Package 1 (ESP 1), which
Shrayer proudly took to a Homebrew meeting,
where he was overwhelmed by the enthusiasm
of his fellow hobbyists.
Although most software at this time was
given away (shared within the community and
thus part of the public domain), people offered
to purchase copies of ESP 1. Shrayer agreed to
charge something, because he had to reconﬁg-
ure the program for each purchaser’s conﬁgura-
tion (microprocessor, available memory, storage
devices, monitor, and printer).
had to provide documentation for each copy of
ESP 1 that he sold. Tired of typing the assembly
listing on a manual typewriter, Shrayer spent a
year creating a program to do this job, unaware
that he was creating a “word processor.”
October–December 2006 33
Figure 1. Michael Shrayer working with Electric
Pencil on a Radio Shack TRS-80 Model III in 1976.
(Photo courtesy of Paul Freiberger.)
By December 1976, after a year of effort,
Electric Pencil (written in 8080 assembler for
the Altair) was finished, and Michael Shrayer
Software was in business.
Because there was
no appreciable retail distribution system for
microcomputer software, Shrayer sold his prod-
ucts through the mail and advertised in hob-
byist magazines like Personal Computing,
Interface Age, and Kilobaud Microcomputer, in
addition to mainline magazines like Byte.
the operator’s manual noted:
Electric Pencil is a Character Oriented Word
Processing System. This means that the entire text
is entered as a continuous string of characters and
is manipulated as such … Since lines are not delin-
eated, any number of characters, words, lines or
paragraphs may be inserted or deleted anywhere
in the text. The entirety of the text shifts and
opens up or closes as needed in full view of the
user. The typing of carriage returns as well as word
hyphenation is not required since each line of text
is formatted automatically … Everything appears
on the video display screen as it occurs thereby
eliminating any guesswork. Text may be reviewed
at will by variable speed scrolling both in the for-
ward and reverse directions.
Formatting and editing used a series of com-
mands, consisting of the Control key (usually
shown as ^) and another key. For example, to
insert characters in a line, the cursor was placed
where the insert should start and the ^F
sequence was entered. Following the comple-
tion of the insert, the insert mode was termi-
nated by another ^F. Blocks of text were
delineated by placing a backslash (\) at either
end of the block. The block could be moved
with ^H or deleted with ^U. Although this
seems cumbersome in retrospect, it was a great
leap forward in 1977.
Electric Pencil also let the user put titles on
each page by typing $ before the title. Page
numbers were handled similarly. As Paul
Freiberger and Michael Swaine commented in
their book, Fire in the Valley: The Making of the
The popularity of Electric Pencil (EP) was so great
that it created a buyer demand that it be on all
microcomputers then available. Shrayer spent
much of his time rewriting the program for dif-
ferent systems. Not only did each kind of com-
puter require a different version, so did each kind
of printer or terminal. Shrayer was constantly
upgrading the Electric Pencil’s capabilities. In all,
he wrote about 78 different versions.
Electric Pencil worked on systems based on
the Intel 8080 and Zilog Z-80 microprocessors,
and it required a minimum of 8K bytes of
memory. Electric Pencil worked with a video
display monitor, a cassette recorder (for stor-
age), and a printer. According to a July 1978
advertisement in Byte, 12 versions were avail-
able to accommodate different conﬁgurations
of video monitors, printers, cassette recorders,
and disk drives.
The success of the Altair and its competi-
tors spawned a growing hobbyist market for
microcomputer systems and peripherals.
However, an operating system was required to
tie the components together—especially disk
storage units. Accordingly, Gary Kildall’s
CP/M slowly became the de facto standard
for all but a few machine lines.
toward this standard was initiated at IMSAI by
Seymour Rubinstein, who as director of mar-
keting inked the first unlimited distribution
contract with Kildall (personal communica-
tion). The existence of an operating system
meant that the programmer could rely on it
to communicate with the disk drive, display
terminal, and printer. Thus, software could be
more easily (and inexpensively) ported to
new hardware configurations.
Electric Pencil II, introduced in 1978, oper-
ated on any 8080/Z80-based microcomputer
running the CP/M operating system. The CP/M
version, available on an 8-inch floppy disk,
required a minimum of 16K bytes of memory
and a video display module with select memo-
ry-mapped video interfaces such as IMSAI’s
VIO, Processor Technology’s VDM-1, and
Polymorphic Computer’s VTI.
recorder, for the storage of ﬁles, was optional.
To accommodate the hardware configura-
tions, eight versions of Electric Pencil II were
In addition to being CP/M compat-
ible, Electric Pencil II supported four disk drives,
simple ﬁle management, and a cassette backup
capability. As a word processor, the new version
introduced dynamic print formatting, line and
paragraph indentation, centering, underlining,
and boldface. The ability to support four disk
drives let users put Electric Pencil in drive A and
34 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
Electric Pencil worked
with a video display monitor,
a cassette recorder (for
storage), and a printer.
October–December 2006 35
a data disk in drive B for a much improved
By 1981, Micro Data Systems of Los Angeles,
California, had introduced the Pencil Sharpener.
As one might guess from its name, the Sharpener
was designed to enhance Electric Pencil, by pro-
viding a form-letter capability with the ability to
create a data ﬁle with names, addresses, and
other pertinent information. The Sharpener
used the Pencil’s search and replace feature to
insert items from the data ﬁle.
The operator’s manual available for “The
Electric Pencil 2.0z,” listed a number of planned
future programs, among them: Blue Pencil, to
read text ﬁles and note any spelling errors; Red
Pencil, to automatically correct errors detected
by Blue Pencil; and Talking Pencil, a word
processor for visually handicapped persons.
Although a version of Electric Pencil was creat-
ed for the IBM PC, the package did not survive
the increasingly competitive marketplace.
Figure 2 shows a 1986 ad for Electric Pencil.
Steven Jobs and Steve Wozniak also were
avid attendees at the Homebrew meetings.
Wozniak designed his Apple I in late 1975, and
after showing it at a Homebrew meeting, Jobs
and Wozniak offered it for sale—in kit form—
for $666 in summer 1976. The Apple II, a com-
plete system, was introduced at the West Coast
Computer Faire in 1977, and offered for sale
the following year.
An early Apple II application was the simple
word processor called EasyWriter, written by
John T. Draper.
Draper (see Figure 3) had cho-
sen the name TexWriter for his creation but
changed it to EasyWriter after a friend, Steve
Sawyer, told him about the cult classic movie
Easy Rider (1969) in which two counterculture
bikers ride from Los Angeles to New Orleans
in search of America. Three things make
EasyWriter interesting: first, it was written
while Draper was in jail; second, it was written
for the Apple II—the first non-kit microcom-
puter—which changed the microcomputer
marketplace forever; and third, it was (later)
selected by IBM as its word processing software
when the company introduced the IBM PC.
Draper was also known as Captain Crunch
for his exploits as a phone “phreaker.” Using a
whistle he had gotten in a box of Cap’n
Crunch cereal (see Figure 4 next page), Draper
could mimic telephone circuitry tones and
make long-distance calls without paying. He
also made “blue boxes”—tone generators—
which could be used to access long distance
lines illegally. While traveling around the coun-
Figure 2. Electric Pencil PC advertisement, PC Magazine, vol. 5, no. 21,
9 Dec. 1986, p. 104.
Figure 3. John Draper programming EasyWriter, June 1980. (Courtesy
of Bill Baker.)
try, Draper gladly explained his gadget to any-
one he encountered. One such person was
Wozniak, who learned about Draper’s exploits
in Esquire magazine.
Wozniak (Woz) and
Draper subsequently became good friends.
In fall 1977, Draper visited Woz at Apple
Computer and expressed an interest in helping
design a digital telephone card for the Apple II.
He was hired and worked with Woz on the
design. Later, when Draper was arrested for
phone phreaking, he had an Apple II computer
with him. Fortuitously, Draper’s attorney was
able to get him into a work-furlough program,
allowing Draper to work outside the jail during
the day. According to Draper, “It was the per-
fect coding environment.”
write code at night while in jail and test it dur-
ing the day on his employer’s computer.
Because EasyWriter was written in assembly
language and Forth, Draper did not need an
Draper believes it was the need to document
his Forth compiler that motivated him to write a
word processor. Recognizing that Electric Pencil
was the word processor of the day, Draper also
contacted Michael Schrayer and asked if he
could use the same dot commands for printer
In another fortuitous circum-
stance, at the fourth West Coast Computer Faire,
in May 1979, the Forth Users Group booth was
next to Apple Pi, the local Apple users group,
where Draper was demonstrating EasyWriter.
According to Draper, “We decided to sell copies
for $69.69 and couldn’t copy the disks fast
enough to meet the demand.”
When Draper was released from jail, he and
a friend, Matthew McIntosh, established Cap’n
Software Inc. to continue development of his
Forth compiler and EasyWriter. They contract-
ed with Information Unlimited Software (IUS)
to market their products. Although EasyWriter
was popular, Draper went to work on a new
version that would display 80 characters rather
than the 40 displayed on the Apple II; he also
wanted to display lowercase letters (the Apple
II displayed only uppercase).
The new version, EasyWriter Professional,
debuted at the Minneapolis Word Processor
show in summer 1980. Other exhibitors were
Wang, DEC, and IBM, displaying the stand-
alone word processing systems discussed in
Thomas Haigh’s article (“Remembering the
Office of the Future: The Origins of Word
Processing and Ofﬁce Automation”) elsewhere
in this issue. Draper recalled:
By that time, we had TRUE proportional spacing
on the Diablo and Qume daisywheel printers.
People at the show practically laughed at our lit-
tle dinky Apple word processing system, until
they saw the super high quality prints it made.
Personal Computing (January 1981) judged
EasyWriter the best of the bunch.
was the only package “that displays text on the
screen exactly as it will appear on the printed
page.” However, the system only worked on an
Apple II with the addition of an 80-column
board, which at that time was a fairly expensive
($300 to $350) addition. Moreover, EasyWriter
used a nonstandard disk operating system that
made its ﬁles incompatible with those created
by other software.
IBM, however, was planning to enter the
fray with a new microcomputer it dubbed the
personal computer or PC. To complete its ini-
tial offering, IBM needed a word processor. IBM
had discussions with Seymour Rubinstein of
MicroPro about porting that company’s
WordStar program to the IBM environment,
but these broke down when IBM demanded
control over the resulting product.
“Recollections: The Rise and Fall of WordStar”
by Rubinstein elsewhere in this issue.) In July
1981, Draper learned that IUS had agreed to port
EasyWriter to the IBM PC. Although Draper
worked on the conversion for six months, the
36 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
Figure 4. A Cap’n Crunch cereal box. (Courtesy
Quaker Oats Co.)
product was not ready for release on time, and
Draper was pressured into agreeing to the release
of an earlier version.
When the IBM PC was released in October
1981, EasyWriter was the word processor, at
$175. The result was a disaster for Draper and
IUS. The author of “Not So Easy Writer,” a review
in the ﬁrst issue of PC Magazine, was critical of
A reviewer for Byte was even more
critical, suggesting that users purchase other
computer systems, such as the Radio Shack TRS-
80, if they couldn’t wait for a better word proces-
sor to be offered for the IBM machine.
Although IBM offered free updates, the product
never fully recovered from such poor reviews.
Perhaps this situation was as it should be.
Industry observer Amy Wohl stated that IBM did
not want a strong word processing program
because it was concerned that such a package
would take sales away from its DisplayWriter
stand-alone word processing system.
“How We Process Words: The Marketing of WP
Software” by Wohl elsewhere in this issue.) The
demise of EasyWriter demonstrated the impor-
tance of good reviews at a time when most pur-
chasers were hobbyists who relied on periodicals
and user groups for their information.
The ﬁrst retail stores
“During the second half of the 1970s, as the
personal computer began to take off, a retail
infrastructure came into being,”
Kelly has noted. One of the ﬁrst retail stores was
situated in California. Dick Heiser, a member of
the Southern California Computer Society, an
early user group,
realized that since most of the
attendees owned Altair kits, they would need
assistance and additional hardware. Heiser
opened Arrow Head Computer Company in Los
Angeles in mid-July 1975. On the store’s sign,
Heiser added “The Computer Store” because he
thought it sounded good. Heiser discovered that
owners of the $395 Altair kits were willing to
spend 10 times that much on additional hard-
ware such as video terminals, disk drives, print-
ers, and additional memory. By November 1975,
John French had opened his Computer Mart
focusing on the IMSAI 8080 computers (Altair
clones), and on 8 December, Paul Terrell opened
his Byte Shop in Mountain View, in the heart of
the Silicon Valley. By January 1976, numerous
entrepreneurs had contacted Terrell and offered
a part of their future proﬁts for the use of the
name and some of his retailing experience.
On the other side of the US, Dick Brown
opened a shop, also called The Computer Store,
on Route 129 in Burlington, Massachusetts.
Elsewhere, entrepreneurs opened individual
stores to sell hardware and the small, but grow-
ing, number of available software packages,
including video games, and applications for
accounting and record keeping. By the end of
June 1976, 235 independent retail stores sold
It should therefore come as no
surprise that two former IMSAI employees
would start the ﬁrst franchise operation. The
ComputerLand retail chain, started by Ed Faber
and Bill Millard, was incorporated on 21
September 1976; the ﬁrst store opened in
Hayward, California, on 10 November 1976.
Other dealers offered computers from Southwest
Technical Products, Polymorphic Systems, and
Processor Technology, among others.
Another major step in the creation of a
viable and robust retail infrastructure occurred
when Tandy Corporation’s Radio Shack divi-
sion announced its TRS-80 microcomputer in
August 1977. This fully assembled machine
retailed for $399, and by September, Tandy had
sold 10,000 TRS-80s and peripherals. What’s
more important, Tandy also sold the software
to justify these purchases. The low-cost, assem-
bled TRS-80s served as the introduction to
computing for tens of thousands of “people
who didn’t need a computer.” By 1981, the
software market alone was estimated at $150
million, and software-only stores started to
Between 1981 and 1983, software sales
grew from $81 million to $495 million.
Along with the growth of the retail market,
with its inevitable advertising in local newspa-
pers as well as nationwide magazines, the
number of microcomputer users exploded, and
an insatiable demand for high-quality soft-
ware—including word processing software—
transformed the fledgling software industry.
According to Campbell-Kelly:
The word processor quickly became an essential
application for every personal computer on the
market. At ﬁrst, it was a fragmented market, with
October–December 2006 37
The low-cost, assembled
TRS-80s served as the
introduction to computing
for tens of thousands of
“people who didn’t
need a computer.”
no dominant product and with each computer
having its own word processing package—
AppleWriter for the Apple II, Scripsit for the TRS-
80, and WordPro for Commodore machines. …
By 1979, however, with the availability of 80-
character-wide screens, it became possible to
offer facilities comparable to those of a profes-
sional word processing system. The most suc-
cessful package by far was MicroPro’s WordStar.
In 1976, Seymour Rubinstein (see Figure 5)
was the director of marketing for IMSAI.
part of his responsibilities, Rubinstein negoti-
ated a number of important licensing agree-
ments that allowed IMSAI to market a more
powerful computer than its competitors. He
negotiated the ﬁrst contract for the distribution
of the CP/M operating system with Gary Kildall
of Digital Research.
Rubinstein also negotiat-
ed a contract with Gordon Eubanks for his C-
Basic compiler, and with Bill Gates and Paul
Allen for a Fortran compiler.
After an argument with IMSAI’s founder and
chairman, Bill Millard, Rubinstein left to start his
own company. Given his experience in the small
but growing software industry, Rubinstein want-
ed to create a professional software business to
compete with entrepreneurs like Shrayer and
Draper. In addition, Rubinstein eschewed selling
to manufacturers or individuals, and focused his
attention on the growing retail infrastructure.
Realizing that he needed an expert programmer,
he sought out Rob Barnaby who had left IMSAI
two weeks before Rubinstein.
John Robbins Barnaby was a systems pro-
grammer who had worked on enhancements
to the CP/M operating system.
with the extensive programs Barnaby wrote to
test Eubanks’ C-Basic compiler, Rubinstein
called him and suggested that they work
together:—Barnaby would do the coding and
Rubinstein would do the marketing of two new
programs: WordMaster and SuperSort. Barnaby
quickly wrote the two programs in 8080 assem-
bly language. WordMaster was similar to soft-
ware that Barnaby had created when he was
with IMSAI. SuperSort resembled utility pro-
grams for the IBM System/360. In a 3 May 2000
email to Mike Petrie, Barnaby stated:
Seymour was the marketing brains—it was he
that said we should address word processing to
get a larger market. The deﬁning change was to
add margins and word wrap. Additional changes
included getting rid of command mode and
adding a print function. I was the technical
brains—I ﬁgured out how to do it, and did it, and
documented it. The product’s success I think
related both to it being the right product
(Seymour) and to it being a fairly good imple-
mentation given the equipment (me).
As it happened, CP/M contained a com-
mand line editor used by programmers to enter
and edit programs. Barnaby developed a way of
showing the editor on a monitor, which made
writing and debugging programs much easier:
It was only approximate because it couldn’t show
things like page breaks and stuff like that. It just
showed the text so it looked like screen paper.
Another problem with the [editor] program was
that it only ran on IMSAI’s version of the CP/M
operating system and it was also missing a lot of
features. I asked him to write a new editor from
scratch with a lot of improvements, so it could
run on the standard CP/M and we could sell it to
Because Rubinstein didn’t have much time
to establish a company, he filed a fictitious
business name statement as MicroPro
International Corporation in September 1978.
Rubinstein chose the name MicroPro to
emphasize his goal of building professional soft-
ware for the growing microcomputer indus-
In those early days of microcomputing,
38 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
Figure 5. Seymour I. Rubinstein in 1984. (Photo
courtesy of Seymour I. Rubinstein.)
this was a rare insight by Rubinstein, and
showed a willingness to gamble his future on
In October 1978, Rubinstein was approached
by the author of a program called TextWriter,
which allowed control characters to be embed-
ded in the text. These control characters pro-
vided instructions to the print program, which
formatted the output for the printer. Rubinstein
agreed to add TextWriter to MicroPro’s offer-
ings; he also decided that the market was ready
for a better-quality word processing package
and obtained a copy of a DataPro report on the
state of word processing. This report identiﬁed
the features offered by the various (minicom-
puter) stand-alone packages being sold by com-
panies such as Lanier, Xerox, IBM, and Wang.
Based on this information, Rubinstein planned
to create a word processing package that incor-
porated the capabilities of minicomputer word
processors and worked with the CP/M operat-
As a result, Barnaby set to work on
an elaboration of WordMaster, which was mar-
keted as WordStar. By April 1979, Barnaby had
coded an alpha version of WordStar. Writer
John Dvorak reported that
In four months, Barnaby wrote 137,000 lines of
bullet-proof assembly-language code. Rubinstein
later checked with some friends at IBM who cal-
culated that Barnaby’s output was equal to 42
man years [of effort].
MicroPro started shipping WordStar in June
1979. After completing most of the code,
Barnaby was joined by Jim Fox, who helped
complete the code and wrote the installation
program. Barnaby and Fox continued to im-
prove WordStar through several interim ver-
A major innovation was the development
of the MailMerge program, which allowed the
insertion of names and addresses into a WordStar
document from a separate ﬁle. According to
Rubinstein (personal communication), Microsoft
still uses the name “Mail Merge to describe this
function without regard to its origin.”
WordStar was very powerful, and it had a
carefully designed Control-key interface,
designed to allow speedy input of text. Indeed,
Rubinstein designed the interface so that the
left hand controlled the cursor’s movement and
the right hand controlled other functions. The
most important feature, however, was that the
text on the screen looked like the printed out-
put. With other word processors, users had to
print the file to see what it looked like, and it
usually took a few passes to get the formatting
right. Thus, WordStar was the ﬁrst “What You
See Is What You Get” word processor, a term
Rubinstein used in advertising the product (see
Although the phrase was used in
other contexts, Rubinstein adopted it to use in
his advertising (personal communication).
According to Rubinstein, WordStar was the
first program to show on-screen page breaks,
have in-line help, and be keystroke sensitive.
WordStar did automatic word-wrap and
hyphenation, and users could set left and right
Indeed, Rubinstein believes that
this WYSIWYG capability of WordStar is what
allowed the Intel platform to capture market
Rubinstein believes that WordStar
“provided a reason for someone coming into a
computer store to buy a machine (because)
WordStar … made the computer functional
and useful immediately.”
October–December 2006 39
Figure 6. WordStar advertisement, Personal Computing, January 1981.
(Courtesy of Seymour I. Rubinstein.)
To Rubinstein, WordStar was the ﬁrst “killer
application” because of its dominance of the
marketplace and its millions of dollars in sales.
As if to back him up, MicroPro’s revenues grew
from $500,000 in 1979 to a high of $72 million
in ﬁscal year 1984, as Table 1 shows.
By anyone’s measures, this was a phenome-
nal pattern of growth, and indicates the sales of
enormous number of copies of WordStar, which
was overwhelming its competition.
not take into account the enormous number of
pirated copies, which Rubinstein estimates at
three times sales (personal communication).
According to editor and publisher Efrem Sigel,
WordStar had 650,000 cumulative unit sales by
fall 1983 with a retail value of $325 million. The
next two word processing programs, PFS-File
and EasyWriter, had 250,000 and 55,000 units,
MicroPro not only had a superi-
or word processing package, but under
Rubinstein’s direction, the company continued
to improve its products, and developed power-
ful marketing programs to promote them.
MicroPro’s advertising expenditures for 1983
were projected at $1,950,000, an amount
exceeded only by the Lotus Development,
Microsoft, and Ashton-Tate corporations.
result was that MicroPro garnered a 10 percent
share of the personal computer software mar-
ket, based on 1983 revenues, behind Microsoft,
VisiCorp, and Lotus Development.
In 1982, Rubinstein established a sales force
that resulted in a quintupling of revenues, a
direct result of WordStar’s being on the software
distributor Softsel’s “Hot List” for 61 weeks.
Given MicroPro’s sales force of, by Rubinstein’s
estimate, 150 people, the revenue explosion is
not surprising. By November 1983, MicroPro had
sales ofﬁces in 20 US and 6 European cities.
Another contribution to the growth in
WordStar sales resulted from the documenta-
tion’s being translated into 42 different lan-
guages. Rubinstein did the ﬁrst translation (to
French) with the help of the staff in the Paris
To expand his empire, Rubinstein also
looked for partnerships with hardware suppli-
ers such as Osborne Computer Corporation, to
which he sold WordStar in return for part own-
ership. MicroPro also licensed WordStar to
other computer manufacturers to package with
their machines for ﬁrst-time buyers.
However, all was not roses at MicroPro.
Growing a company is always a difficult
process, but accommodating and institutional-
izing rapid growth in personnel is not without
its problems. Rob Barnaby had been working
long hours and felt burned out. In October
1979, he took a holiday and didn’t return. In
January 1980, Rubinstein called Barnaby and
asked if he was ready to return, offering back
pay and reinstatement of Barnaby’s stock
options. When this did not work, Rubinstein
gave him additional time, yet Barnaby and
Rubinstein ultimately parted ways in March
MicroPro also added numerous new
employees in its development, marketing, and
other departments. In 1981, to gain additional
capital, shares were sold to Fred Adler, a ven-
ture capitalist. Because the company’s fortunes
changed so radically—for the worse—after-
ward, Rubinstein was later to comment that
this was , “the biggest mistake of my life.”
In 1984, just in time for the initial public
offering (IPO), sales grew to $70 million, and
MicroPro was the world’s largest software com-
pany. On 7 January 1984, two months before
the public offering, Rubinstein suffered a heart
attack. While he was still in the intensive care
unit, Rubinstein’s attorneys told him that his
presence on the management team would
harm the IPO. He signed the papers they had
brought, which gave him the title of Chairman
Emeritus, and changed his stock to nonvoting.
In exchange, he received $7.5 million. H. Glen
Haney, formerly of Sperry Univac, took over
management of MicroPro and within the next
few months, Rubinstein was gradually eased
out of any role in the company.
According to John Dvorak, “from that
moment on, the company lost its edge. Haney,
an ex-Sperry-Univac guy, took over as CEO at
To make matters worse,
Rubinstein claims that WordStar was not adver-
tised for the next three years after he was released
40 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
Table 1. MicroPro revenues in millions of dollars.
Fiscal Year Fiscal Year Fiscal Year Fiscal Year Fiscal Year Fiscal Year Fiscal Year
Parameter 1979 1980 1981 1982 1983 1984 1985
Revenues 0.5 1.8 5.2 25 45 72 40
Annual increase (%) 360 288 480 180 160 −55
* Note: Data extracted from S. Rubinstein, “Recollections: The Rise and Fall of WordStar” article elsewhere in this issue and veriﬁed
with other sources.
from the hospital, resulting in an inevitable
decline in the company’s sales and image.
months later, the stock dropped from its initial
price of $10.50 a share to less than $5 a share.
Rubinstein’s original vision for MicroPro was
to bring out a suite of compatible programs. In
the early 1980s, MicroPro released a spreadsheet
called CalcStar and a database program called
DataStar, and integrated them with WordStar in
a system appropriately called StarBurst.
According to Dvorak, Starburst “was phenome-
nal for its era. In fact, it was the original ofﬁce
suite. Adler and company killed the idea.”
Other problems associated with rapid
growth also plagued the company. Rubinstein
had problems with other members of his devel-
Some of them left and built
startups to compete with MicroPro. Peter
Meraux, Stan Reynolds (MicroPro’s head of
engineering), and Richard Post founded
NewStar. In one of their first contracts, they
agreed to write a word processor for George
Morrow of Morrow Designs, a manufacturer of
In return for start-up capital, Morrow received
a license to use their WordStar clone on his
computers. NewStar released NewWord 1.0 for
MS-DOS in 1983. NewWord 2.0 for MS-DOS,
released in 1984, added numerous features that
WordStar users had been asking for, such as an
un-erase feature, a built-in spelling check fea-
ture, and support for laser printers.
3.0, which added macros, math capability, spell
checking during editing, and support for the
DOS directory (and path support which was
missing from WordStar), was released in 1986.
As new word processing packages entered
the marketplace, the pressure to stay at the top
of the heap intensified. By this time, the soft-
ware industry had settled into a pattern in
which new releases were announced every 12
to 18 months. Unfortunately for MicroPro, per-
sonnel and development problems kept it from
announcing a new release during 1984 and
1985, during which time numerous packages
eclipsed the once-dominant WordStar.
Another strategic error, according to
Rubinstein, was that MicroPro stopped doing
business with the 1,500+ retailers in its network
and started using software distributors, among
them Ingram, Micro, and First software.
According to Rubinstein,
It was a deliberate but disastrous decision … our
loyal dealers were cut off so they had to buy
from the big distributors: MicroPro just simply
stopped doing business with everyone; even
ComputerLand had to buy from the distributors.
The reason given for the shift to distributors
was the problems associated with handling large
order quantities; rather than improving the
order handling process, MicroPro management
chose to change its successful distribution strat-
egy. Rubinstein believes this was a major factor
in the loss of revenues during ﬁscal year 1984.
Because the original WordStar code had been
written in 8080 assembly language, it was difﬁ-
cult to transition the code to other operating
environments. At that time, AT&T was planning
to enter the personal computer market and
needed applications. MicroPro wanted to port
WordStar to the Unix environment. Instead,
Rubinstein purchased a WordStar clone, written
in the C programming language, which ran
Rubinstein also hired the pro-
grammer, Edward de Jong, and told him to hire
a small group of programmers to help him. This
new project was code-named “Ivan,” which was
Rubinstein’s middle name. Later, it was decided
that the project should also produce a version
for the new IBM PC. The Ivan team worked out-
side of normal MicroPro channels, and this was
to have disastrous consequences when the prod-
uct was marketed.
The new version, based on de Jong’s pro-
gram, was named WordStar 2000. However,
WordStar 2000 was slow, used an entirely new
mnemonic-based command structure, and pro-
duced a ﬁle format incompatible with WordStar.
Thus, existing WordStar users had to practically
start over again—if they “upgraded” to WordStar
2000. Although MicroPro supplied a utility pro-
gram to translate ﬁles between WordStar and
WordStar 2000, this lack of upward mobility
meant that WordStar users had to translate all
their ﬁles to use the new software.
Although many software companies experi-
enced technical and other difficulties in this
period, the leading supplier always received
more attention, and WordStar 2000 suffered in
reviews. One reviewer opined that
October–December 2006 41
As new word processing
packages entered the
marketplace, the pressure
to stay at the top of the
WordStar is one of those phenomena, like mod-
ern art, that brings out the critic in people.
Whether they love it or hate it, people will tell
you exactly what’s good or bad about the pro-
gram. But even WordStar loyalists, who blithely
overlook the program’s cluttered menus and
arcane commands, are beginning to clamor for
such state-of-the-art features such as automatic
paragraph reformatting, proportional spacing,
keyboard macros, and multiple window editing.
As if this weren’t enough, existing users and
potential users faced a dilemma: “Which
WordStar should I use?” WordStar 2000 was
priced at $495—the same as WordStar 3.3. The
former contained some added features but
required users to essentially learn a new pack-
age and translate all their ﬁles to a new format.
Most existing WordStar users preferred to stick
with what they were comfortable with, and it
was mostly new users who chose WordStar
2000. Rick Chapman, who had moved to
MicroPro’s marketing department, refers to this
as an error of positioning:
For example, it is stupid to create two products
with the same name, price point, functionality,
and target audience and attempt to sell them at
the same time. This may seem stunningly obvi-
ous, but somehow one of the world’s largest soft-
ware companies, MicroPro, publisher of
WordStar, a product that once ruled the word
processing market, did exactly that.
All of this was occurring at a time when new
word processing packages were being intro-
duced on a monthly basis. In addition,
NewStar’s NewWord package was eating away
at WordStar’s installed base.
Within the (now) MicroPro International
Corporation, the inevitable split of program-
mers and support personnel between “us” and
“them”—the WordStar and the WordStar 2000
teams (and their internal supporters)—caused
additional problems. Rubinstein’s employment
contract with MicroPro expired on 1 April 1986.
The new president, Leon Williams, put his trust
in a new product management team, which
included Rick Chapman. Chapman believes
that the company had hit a new low: “By 1987,
as MicroPro wrestled itself to the mat, it had lost
its leadership of the word-processing market to
Microsoft Word and WordPerfect.”
Given responsibility for straightening things
out, Chapman saw his major tasks as getting an
upgrade for WordStar to market, and resolving
the conﬂict between WordStar and WordStar
2000. To solve problem one, Williams purchased
NewStar for $2 million and used its NewWord
package as the basis for the upgrade. WordStar
4.0 was released in 1987, more than four years
after WordStar for PC DOS, and sold well to
existing users. Reviews, however, were mixed.
The more difﬁcult decision was what to do
with WordStar 2000, which, by this time, had a
signiﬁcant installed base. Chapman’s dilemma
was how to upgrade the package without losing
present users. He decided to “re-christen”
WordStar 2000 as a “word publisher,” which the
MicroPro marketing department deﬁned as “a
word processor with exceptional laser-printing
capabilities, a particular strength of WordStar
2000 at that juncture.”
Ultimately, Chapman was given responsibil-
ity for the product management of the newly
resurgent WordStar. While overseeing the devel-
opment of WordStar 5.0 (to be released in early
1988), he encountered a significant, and per-
haps fatal, problem: the package lacked a print-
er database—the print drivers that translate the
format commands of any application into the
unique commands required by the various
makes and models of printers. Chapman knew
that MicroPro had developed a high-quality
database for more than 300 printers, as part of
the WordStar 2000 effort, and that the compa-
ny had decided to use this database for all future
But, inexplicably, this database was not
incorporated into WordStar 5.0: the developers
had discarded the WordStar 2000 database and
were trying to replace it with a new one based
on hierarchical database technology. Chapman
believes that “It was an incredibly foolish thing
to do and it sealed MicroPro’s fate.”
of discarding MicroPro’s proven printer tech-
nology delayed the release of WordStar 5.0 for
six months, and when it shipped in late 1988,
it was not only behind its competition but it
supported only one-third of the previously sup-
42 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
“By 1987, as MicroPro
wrestled itself to the mat, it
had lost its leadership of the
word-processing market to
Microsoft Word and
October–December 2006 43
Other management decisions added to
MicroPro’s woes. Because WordStar was avail-
able for so many different computers and con-
figurations (and in 42 different languages), it
may have been one of the most copied (pirat-
ed) software packages of all time. In an attempt
to minimize the pirating of its software,
MicroPro released WordStar 2000 on copy-pro-
tected diskettes. To run WordStar 2000, the user
had to swap disks in drive A and drive B sever-
al times. Although MicroPro was not the only
software publisher to adopt copy protection,
the WordStar user community (and software
reviewers) had another grievance against the
As if it didn’t have enough problems, IMSAI
sued MicroPro in 1985, alleging that Rob
Barnaby used code from IMSAI’s New Editor
(NED) project in his development of
WordMaster. While working at IMSAI, Barnaby
had begun coding NED because he didn’t like
the line editor supplied with CP/M. Although
Barnaby and Rubinstein denied the allegations,
IMSAI claimed that the work that Barnaby had
done while at IMSAI belonged to them and so
did WordMaster. A court-ordered discovery
petition uncovered a note written by Millard
five years earlier that said: “sue Rubinstein.”
Since the statute of limitations had expired,
that is, more than ﬁve years had expired since
the discovery of the alleged code infringement,
the case was dismissed.
1984: A market shift
By 1984, IBM’s PC—and the IBM clones that
followed it into the marketplace—had started
a transformation of the marketplace: from one
consisting of hobbyists to one that included
professionals in large and small organizations.
Although introduced in 1981, the IBM PC did
not achieve 50 percent of new purchases until
1986 and did not constitute more than half of
the installed base until 1988.
acceptance of the IBM PC and the MS-DOS
environment, however, resulted in a more
homogenized market, in which an increasing
number of packages competed for the same
buyers. As a result of this shift, and the expens-
es and problems of porting software packages
to the PC/MS DOS environment, there was a
shakeout in the personal computer industry:
No one expected the halcyon days of the person-
al computer software business to pass so quickly.
Industry experts had projected that this market
would continue to double annually, and 3,000
hopefuls, as a result, had jumped into the fray. But
the glut of suppliers, along with the soaring cost
of marketing new products and a ﬂood of me-too
programs, is changing the picture dramatically …
At last count, there were 200 or more word proces-
sors, 150 spreadsheets, 200 database programs,
and 95 integrated packages that offer at least three
functions. Moreover, distributors report that of
the 20,000 programs on the market, a mere 20
make up as much as half of their total business.
The shakeout affected the entire microcom-
puter industry. Numerous hardware manufac-
turers, after struggling for many years, either
began making IBM PC clones or got out of the
business. And, the changing hardware market
had an inevitable negative impact on the soft-
Of the nine largest ﬁrms of 1983 … ﬁve were in
a terminal condition by the summer of 1984.
VisiCorp and MicroPro, publishers of the leading
spreadsheet and word processor had lost market
share to Lotus Development and the WordPerfect
Corporation, respectively. Digital Research had
lost operating system sales to Microsoft.
MicroPro’s revenues, which peaked in 1984
at $72 million, dropped to $40 million for
1985, a devastating drop of 45 percent. Table 2
shows the changes in market shares for word
processing software vendors.
The sidebar, “Word Processing Market in
1985,” more clearly depicts the word process-
ing software market in 1985.
In the late 1970s, thousands of people were
writing microcomputer software, and by the
early 1980s, the marketplace offered hundreds
of word processing packages—written for the
different hardware/software combinations then
available. When the CP/M operating system
became a de facto industry standard, however,
MicroPro’s WordStar became the dominant
Table 2. Word processing market shares in 1984.
Vendor Product Share and Trend
MicroPro WordStar 23% falling
Sorcim SuperWriter 15% rising
Software Publishing PFS:Write 9% falling
MultiMate International MultiMate 8% rising
Microsoft Word 8% rising
Perfect Software Perfect Writer 7 % falling
Information Unlimited EasyWriter 6% falling
Life Tree Software Volkswriter 1% falling
Satellite Software Word Perfect 1% falling
44 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
By mid-1985, the market for word processing software
had exploded. The need for information on the various
word processing packages available was becoming critical
as well. No one could possibly test more than a few pro-
grams at their local computer store, and users looked to
the technical press for guidance. The 28 January 1986 issue
of PC Magazine contained an extensive analysis and test-
ing of 57 word processing packages:
No class of PC software is bigger or more widely used than
word processors. A word processor is every user’s daily com-
puter tool. Thus, more word processors are in use on PCs, by
a variety of users, than any other type of software product.
The number and variety of products designed to fulﬁll this
incredibly wide range of needs has made word processing the
most competitive market in PC software. As a result, no one
product dominates that market the way Lotus Development
Corp.’s 1-2-3 holds sway in spreadsheets or Ashton-Tate’s
dBase product line dominates database management.
Indeed, the PC Magazine author deﬁned seven cate-
gories, based on their target markets. The ﬁrst three cate-
gories (corporate, professional, and personal), included
traditional word processors, that is, those used exclusively as
general-purpose document editors, formatters, and printers.
The other four categories comprise specialty word
processors, which are designed for unique applications or
require special operating environments.
Each section contained an overview graphic of pack-
ages examined and then provided detailed information on
each package. Table A lists the packages that PC Magazine
This material demonstrates what the software market-
place was like in mid-1985, less than four years after the
introduction of the IBM PC. At this time, the marketplace
still demonstrated strong sales of word processing software
for machines using the CP/M, Apple DOS, Macintosh,
Unix, and other proprietary operating systems.
References and notes
1. J. Dickinson, “The Business of Words: Corporate, Professional,
Personal: PC Magazine Labs Takes a Fresh Approach to Testing
Word Processors—The Largest, Most Complex, and Most
Competitive PC Software Market,” PC Magazine, vol. 5, no. 2,
1986, pp. 93-251.
2. Reviews of 20 scientiﬁc, integrated packages, those with spe-
cial operating system extensions, and outline processors
appeared in later issues. The 25 February 1986 issue ( vol. 5,
no. 2), contained “The Business of Words: Special Purpose”
by John Dickinson and examined 12 word processors in
three categories: integrated programs, scientiﬁc, and special
operating systems (pp. 177 to 214). The ﬁnal part of this
analysis, which looked at seven outliners, appeared in the 25
March 1986 issue (vol. 5, no. 6) as “Outliners: A New Type of
Word Processor Tries to Help Us Organize Our Thoughts and
Ideas by Providing a Variety of Logical Structures to Fit them
Into” (pp. 199-220). This series of reviews examined 77
packages and covered 219 pages.
Word Processing Market in 1985
Table A. Word processing packages tested by PC Magazine in 1986. Bolded entries were PC Magazine editor’s choices
based on price and performance.
Corporate Price Professional Price Personal Price
Allegory 1.2 $170 Executive Footnoter 5 $99.95 Bank Street Writer $79.95
Benchmark 4.4 $395 Executive Secretary V $69.95 Cut and Paste $34.94
DisplayWrite 2 1.10 $385 Executive Writer 1.0 $69.95 DisplayWrite 1 1.0 $99
DisplayWrite 3 1.00 $450 Final Draft 3 $395 Easy 1.00 $150
Leading Edge 1.3 $250 The FinalWord 2.0 $395 EinsteinWriter 7.3 $170
MASS-11 5-C $995 Microsoft Word 2.0 $395 FriendlyWriter 3.2 $89.95
MultiMate 3.3 Pro 3.31 $495 My Word! 1.71 $60 Homeword, Pro Ed $69.95
MultiMate Advantage 3.50 $595 NewWord 2.15 $249 Just Write 1.01 $145
OfﬁceWriter 3.5 $445 NewWord 3 3.00 $349 MindReader 2.0 $189
PeachText 5000 2.11 $295 PC-Write 2.55 $75 MyWrite 1.04 $49.95
Professional QWERTY 3H $149 Perfect Writer 2.0 $199 Paperback Writer 1.0 $39.95
SAMNA Word III 3.0 $495 Personal WordPerfect 4.0 $195 PC-Type 1.0 $59.95
Word Result 2.0 $495 PractiWord 1.05 $99 Personal QWERTY 3H $99
WordMARC 4.11 $295 The Professional Writers Package 2.0 $490 PFS:Write C $140
WordStar 2000 Plus 1.01 $595 The Smart WP 2.00 $195 Pronto! 1.0 $69.95
Spellbinder 5.40 $495 TEXTRA 3.1A $24.95
SuperWriter 1.03 $295 WordPerfect jr 4.0 $89
Volkswriter Deluxe 2.2 $295 Wordvision $50
WordPerfect 4.1 $495 Write ’N Spell $149.95
WordStar Professional 3.31 $495 Writing Assistant 1.01 $149
XyWrite II Plus 2.0 $295 ZenWord 1.00 $29.95
word processing package, estimated to have as
much as 80 percent of the market. This was the
start of what economists Stan Liebowitz and
Stephen Margolis call a serial monopoly, in
which WordStar was the market leader until
1986 when it was eclipsed by WordPerfect.
This history continues in the companion arti-
cle, “Proliferation and Consolidation of the
Word Processing Market: 1985−1995,” which dis-
cusses the origin and evolution of WordPerfect
and Microsoft Word, and continues with the
impact of Microsoft Windows on the word pro-
cessing software market.
References and notes
1. The best history of personal computing is
contained in P. Freiberger and M. Swaine, Fire in
the Valley: The Making of the Personal Computer,
2nd ed., McGraw-Hill, 2000.
2. The ﬁnest discussion of application software evo-
lution will be found in M. Campbell-Kelly, From
Airline Reservations to Sonic the Hedgehog: A Histo-
ry of the Software Industry, MIT Press, 2003.
3. S.J. Liebowitz and S.E. Margolis, Winners, Losers
and Microsoft: Competition and Antitrust in High
Technology, The Independent Inst.,1999, p. 10.
4. The computer was named for the technical editor
of Popular Electronics magazine, Les Solomon,
who was a strong supporter of the Processor
Technology efforts to design a computer.
5. M. Campbell-Kelly, “Not Only Microsoft: The
Maturing of the Personal Computer Software
Industry, 1982−1995,” Business History Rev., no.
75, spring 2001, p. 127.
6. A strict interpretation of Annals “15 year rule”
would prohibit discussion of events after 1991;
however, a more convenient ending has been
chosen based on events surrounding Microsoft
Word for Windows.
7. P. Freiberger and M. Swaine, Fire in the Valley: The
Making of the Personal Computer, pp. 41-49.
8. Ibid., p. 149.
9. M. Campbell-Kelly, From Airline Reservations to
Sonic the Hedgehog: A History of the Software
Industry, MIT Press, 2003. p. 217.
10. The Homebrew was an early user “information
exchange” between people interested in
computers. The ﬁrst meeting of the Amateur
Computer Users Group (aka Homebrew Comput-
er Club) took place on 5 March 1975 in Gordon
French’s garage in Menlo Park, Calif.
11. Bob Marsh and Gary Ingram founded Processor
Technology to manufacture microcomputers in
1975; the other manufacturers at this time were
Micro Instrumentation Telemetry Systems
(MITS), which built the Altair, and Cromemco,
founded by Harry Garland and Roger Melen.
12. P. Freiberger and M. Swaine, Fire in the Valley, p. 146.
13. See “Electric Pencil,” in P. Freiberger and M.
Swaine, Fire in the Valley, pp. 186-192.
14. For an insight into Michael Shrayer, see “Confes-
sions of a Naked Programmer (Reminiscence—
Software, Stores and Magazines),” Creative
Computing, vol.10, no. 11, 1984, pp. 130ff;
15. P. Freiberger and M. Swaine, Fire in the Valley, p.
16. See, for example, Byte, vol. 3, no. 7, 1978, p. 77.
17. The Electric Pencil Word Processor, operator’s man-
ual, copyright 1977 by Michael Shrayer, can be
found on Ira Goldklang’s site: http://www.trs-80.
com. The operator’s manual is in upper and lower
case; printing was done on a Diablo 1620 using
an OCR-B printwheel and a carbon ﬁlm ribbon.
18. For example, the Apple and Radio Shack
machines had proprietary operating systems; the
IMSAI, Cromemco, Kaypro, and Osborne
machines had a version of CP/M.
19. The periodical literature uses Electric Pencil II,
although the title on the operator’s manual is The
Electric Pencil2.0z; see http://www.trs-80.com.
20. A.R. Miller, “The Electric Pencil for CP/M,” Inter-
face Age, vol. 3, no. 8, 1978, p. 148. Additional
reviews of Electric Pencil can be found in J.A.
Greenleaf, “Michael Shrayer’s Electric Pencil,”
Personal Computing, vol. 11, no. 5, 1979, pp. 73-
74; and R. Hallen, “Super Word Processors,”
Microcomputing, June 1980, pp. 214-217.
21. Electric Pencil II advertisement, Byte, vol 3, no. 7,
1978, p. 77.
22. R. Hallen, “Super Word Processors,” p. 215,
opined that “Just about any text-editing facility
can be used to create the data list . . .” Another
unusual feature of the Sharpener is that it did not
come with an instruction manual; the master disk
contained some information ﬁles that could be
read on the screen or printed out. It was not for
the faint of heart!
23. M. Shrayer, The Electric Pencil 2.0z Operators Man-
ual, appendix IV, “Future Electric Pencil
Products,” Oct. 1981, pp. 114-115. This manual
was prepared using Electric Pencil 2.0 on a TRS-
80 Model I (48K RAM and 4 disk drives), and an
NEC (probably a daisy wheel) printer operating
under the NEWDOS/80 2.0 operating system;
prooﬁng and spelling corrections were made
with pre-release versions of the Blue Pencil and
Red Pencil dictionary/correction programs from
24. P. Freiberger and M. Swaine, Fire in the Valley, p. 291.
25. Esquire, Oct. 1971, pp. 116ff. Further insight into
this interesting period can be found in A. Lundell
and G.M. Haugen, “The Merry Pranksters of
Microcomputing,” at http://www.atariarchives.
php and “The Ofﬁcial Phreaker’s Manual by Ron
October–December 2006 45
46 IEEE Annals of the History of Computing
The Origins of Word Processing Software: 1976−1985
Rosenbaum at http://www.webcrunchers.com/
26. J.T. Draper, “The Creation of EasyWriter”;
story/home.html. (Draper’s homepage is at
27. Forth is unique among programming languages
in that its development and proliferation has
been a grassroots effort of professional applica-
tions, developers, and hobbyists. See E.D. Rather
et al., “The Evolution of Forth,” History of
Programming Languages, T.J. Bergin and R.G. Gib-
son, eds., ACM Press, 1996, pp. 625-670.
28. J.T. Draper, “The Creation of EasyWriter”; http://
29. This was at a time when most software was given
away through the sharing programs of the various
user groups for Apple, Radio Shack, Commodore,
and other machines; thus EasyWriter and Electric
Pencil would have been among the earliest
software to be sold rather than shared.
30. S. Jong, “Word Processing Software Roundup,”
Personal Computing, Jan. 1981, pp. 26-33.
31. P. Freiberger and M. Swaine, Fire in the Valley, p.
32. P. Freiberger and M. Swaine, Fire in the Valley, pp.
33. A. Fluegelman, “Not So Easy Writer,” PC, vol. 1,
no. 1, 1982, pp. 35-39.
34. G. Williams, “A Closer Look at the IBM Personal
Computer,” Byte, vol. 7. no. 1, p. 62 (full review
is pp. 36-68).
35. P. Freiberger and M. Swaine, Fire in the Valley, p.
36. A. Wohl, interview by T. Bergin, 2 Oct. 2005.
37. M. Campbell-Kelly, From Airline Reservations to
Sonic the Hedgehog, p. 209.
38. This discussion is based on P. Freiberger and M.
Swaine, pp. 231-249.
39. S. Veit, “The Computer Store Saga,” Creative
Computing, vol. 10, no. 11, 1984, p. 135.
40. P. Freiberger and M. Swaine, Fire in the Valley, p.
41. E. Sigel and the staff of Communications Trends
Inc., Business/Professional Microcomputer Software
1986, Table 2.2, “Estimated Rev-
enues for Leading Business/Professional Microcom-
puter Software Publishers and Distributors,” p. 19.
42. M. Campbell-Kelly, From Airline Reservations to
Sonic the Hedgehog, p. 217.
43. For an interesting, short biography of Rubinstein,
see R. Levering, M. Katz, and M. Moskowitz, The
Computer Entrepreneurs: Who’s Making It Big and
How in America’s Upstart Industry, New American
Library, 1984, pp. 214-220.
44. W. Rubinstein, “Recollections: The Rise and Fall of
WordStar,” IEEE Annals of the History of Comput-
ing, vol. 28, no. 4, pp. 64-72.
45. According to M. Petrie, “A Potted History of
46. M. Petrie, “A Potted History of WordStar,” Word-
Star Is Born section; http://www.wordstar.org/
47. See E. Myers, “Wanted: Software for Micros: The
Success or Failure of Microcomputer Companies
in This Decade Will Be Determined by Software,”
Datamation, vol. 27, no. 8, 1981, pp. 56-60.
48. Indeed, Rubinstein took no salary during the ﬁrst
nine months of MicroPro’s existence.
49. J.C. Dvorak, “Whatever Happened to WordStar?”
Computer Shopper, July 1998; see http://www.
50. M. Petrie, “A Potted History,” WordStar is Born
51. WordStar advertisement, Personal Computing, vol.
5, no. 1, 1981, p. 9. In comments on WordStar,
Rick Chapman reminds us that “WYSIWYG” in
1978 did not mean what it means today in a
Windows environment (see M.R. Chapman, In
Search of Stupidity: Over 20 Years of High-Tech
Marketing Disasters, Apress, 2003, p. 49).
52. At this time, most users had to print their materi-
al many times to see the effect of formatting
commands, which were embedded in the text
but not shown on the screen; some later
packages, such as WordPerfect, let users toggle
between versions of a document displaying or
suppressing such format commands.
53. Waldrop and other authors believe WYSIWYG was
introduced by the Bravo text editor at Xerox
PARC; Wikipedia attributes “WYSIWYG” to a
newsletter published by Arlene and Jose Ramos,
called WYSIWYG, which was created for the
emerging prepress industry going electronic in the
late 1970s. Some authors believe that the origin of
WYSIWYG was a re-use of a popular line used by
comedian Flip Wilson doing “Geraldine,” a charac-
ter on the Flip Wilson Show in the early 1970s.
54. A “killer application” is an application that is so
compelling that it motivates people to purchase
a computer to be able to use the application.
Although numerous sources identify VisiCalc as
the ﬁrst killer app, I believe that Rubinstein’s
assertion is at least as compelling.
55. In 1984, WordStar had 23 percent of the word pro-
cessing market. See R.T. Fertig, The Software Revolu-
tion: Trends, Players, Market Dynamics in Personal
Computer Software, North-Holland, 1985, p. 164.
56. E. Sigel and the staff of Communications Trends
Inc., Business/Professional Microcomputer Software
1986, Table 3.3, “Best Selling
Applications Software Programs,” fall 1983, p.
38. Sigel’s Table 3.6, “Representative Word Pro-
cessing Programs for the IBM PC,” Oct. 1983,
listed 30 programs from 30 publishers and the
comment that “this list by no means covers all
the word processing programs available for the
IBM PC.” (p.40).
57. E. Sigel, Business/Professional Microcomputer Soft-
ware Market, Table 4.5, “Nine-Month and
Projected 1983 Space Advertising Expenditures
of Leading Business/Professional Software Com-
panies,” p. 55.
58. E. Sigel, Business/Professional Microcomputer Soft-
ware Market, Table 6.1, “Revenues and Share of
Market for Leading Business/Professional Micro-
computer Software Publishers,” p. 74.
59. E. Sigel, Business/Professional Microcomputer Soft-
ware Market, MicroPro Corporate Summary, p.
120. This summary listed Rubinstein as chairman,
Glen Haney as president and chief executive ofﬁ-
cer, Frank P. Frost as vice president of domestic
sales, and William G. Crowell as vice president of
product management and development.
60. Quoted in J. Dvorak, “Whatever Happened to
61. R. Levering et al., The Computer Entrepreneurs, p.
62. M.R. Chapman, In Search of Stupidity, p. 50.
63. M. Petrie, “ A Potted History,” WordStar for Win-
dows section; http://www.wordstar.org/
64. M. Petrie, “A Potted History,” Rise of [a] NewStar
65. T. Datz, “A Brighter Star?” PC World, vol. 3, no. 4,
1985, pp. 116-126.
66. M.R. Chapman, In Search Of Stupidity, p. 10.
67. Ibid., p. 55.
68. M.J. Miller, “First Look: Long-Awaited. WordStar
Update Brings Program on Par with Competitors,”
InfoWorld, vol. 9, no. 7, 16 Feb. 1987, p. 47.
69. Ibid., p. 56.
70. Ibid., p. 58.
71. M. Petrie, “A Potted History,” MicroPro Sued over
WordMaster section; http://www.wordstar.org/
72. J. Steffens, Newgames: Strategic Competition in the
PC Revolution, Pergamon Press, 1994, pp. 210-
211. Steffens’ data shows the Intel percentage
increasing to 76 percent by 1992 (where his data
73. “The Shakeout in Software: It’s Already Here,”
Business Week, 23 Aug. 1984, pp. 96-98.
74. M. Campbell-Kelly, “Not Only Microsoft: The
Maturing of the Personal Computer Software
Industry, 1982−1995,” Business History Rev., vol.
75, spring 2001, p. 108.
75. R.T. Fertig, The Software Revolution: Trends, Play-
ers, Market Dynamics in Personal Computer
Software, North-Holland, 1985, p. 164. (To sim-
plify Table 2, I deleted some packages.)
76. S. Liebowitz and S. Margolis, p. xi.
Thomas J. (Tim) Bergin is a
professor emeritus of computer
science and information sys-
tems at American University,
retiring in 2002. In 1965, he
began his career as a computer
systems analyst for the US
Veterans Administration. In
1982, he joined the Center for Technology and
Administration’s faculty. A former editor in chief of
Annals, he currently serves Annals as a senior consult-
ing editor. Bergin has a BA in English from the
University of Maryland, and an MA and a PhD in pub-
lic administration from American University. Most
recently, he edited 50 Years of Army Computing: From
ENIAC to MSRC (Army Research Laboratory, 2000).
Readers may contact Tim Bergin about this article
October–December 2006 47