The Rise, Fall, and Surprising Survival of WordPerfect

The story of the former king of word processors, its battle with Microsoft, and its surprisingly lengthy survival into the present day

Jan 15, 2023

As the personal computer market exploded in popularity in the late 1970s, many people bought them first, and then tried to figure out what useful purpose they could possibly be used for. Early computers such as the Apple II, TRS-80, and Commodore PET were sharply limited machines, especially in comparison with their own later revisions. The differences primarily come down to the amount of available RAM, which was the major bottleneck for software.

For example, the Apple II jumped from a base configuration of only 4k of RAM in 1977, to a full 64k(Apple IIe) just a few years later in 1983. The other two legendary microcomputers from 1977, the Commodore PET and the TRS-80, were similarly equipped with meager RAM amounts when they launched. And even this miniscule amount of RAM was a significant improvement over original microcomputers such the Altair 8800, or one of its clones such as the IMSAI 8080, which might come with as little as 256 BYTES (not a typo) of RAM.

These sharply limited machines strained the imaginations and talent of their owners to find a proper use for them after the initial flush of excitement had worn off. Advertisements in this era not infrequently strained credulity a bit with their descriptions of why you should buy a personal computer…such as suggesting it as a great way to store recipes.

Because as we all know, booting up an Apple II, then inserting the disk with the desired program, then finding the disk with the correct recipe on it and inserting it, typing the DOS command to load the disk’s contents, and then loading the desired receipt…was far superior to merely sorting through a recipe box of neatly typed/handwritten 3x5 cards.

A legitimate early use of personal computers was to replace the typewriter, as the modest requirements of basic word processing lent themselves well to the equally modest capabilities of these early machines. By the early 1980s, the word processing market was dominated by two main rivals: WordPerfect and WordStar. Microsoft wasn’t even in the running yet, as the initial release of what would become Microsoft Word in 1983 came and went with hardly a splash outside of the tiny world of Macintosh software. However as we all know, eventually Microsoft would definitely win the word processing wars, and Microsoft Word would become the de facto standard word processor across the world, without a serious rival.

Yet although WordStar has long since gone defunct, with its only current relevance being the trivia answer that George R R Martin still uses 1987’s WordStar 4.0 to do his writing, WordPerfect still soldiers on, with boxed copies still available at stores such as Best Buy and Office Depot. A faded remnant of what used to be the dominant word processor for MS-DOS, somehow still clinging on to enough market share to be updated and sold.

This is the story of how WordPerfect came into existence, rose to dominance, descended into irrelevance, yet still has managed to somehow hang on when so many of Microsoft's former competitors are now gone and forgotten.

In 1977 Alan Ashton, a computer science professor at BYU created a basic word processing program while on his summer vacation.1 Created more as an interesting way to pass his summer vacation than for any commercial purpose, Ashton decided to create a word processor that avoided a lot of headaches of poor user experience that contemporary word processors inflicted on their users. He already had some experience with text processing from the standpoint of the early text editors used for writing code and had also done some consulting work for the Hill Air Force Base on text processing.

Word processors of that era were clunky things that behaved very differently from what we would expect today, typically being what were called “run-off systems.” For example they usually had visible formatting codes sprinkled among the actual text, such as .LM12 or @HD@B, codes that told the printer how to format the text. Equally annoying was the fact that most word processors did not show line breaks and page breaks on screen as they would print, or what would later be termed WYSIWYG (What You See Is What You Get).

This also had the effect of requiring an additional step prior to printing called repagination, when the computer took the document, formatted it according to the embedded codes, renumbered the pages to be 80 columns wide (most computers of that era tended to have only 40 column wide displays without additional hardware) and then sent finally it to the printer. This step introduced a perceptible wait on the slow computers of the 1970s, in addition to forcing the user to print out the document in order to actually see how it would look.

Additionally, viewing a text document typically had to be done on a page-by-page basis, you could not scroll a line at a time. Finally, inserting text into an existing block not infrequently erased the old text as the user typed forward, as inserting text into an existing block was beyond the capabilities of those very early word processors. This was still a vast improvement over using a typewriter…but when you add the cost of a dedicated word processor, a cost that could run into the five figure range…most people chose to stick with their trusty typewriter.

In what was probably a legacy from Xerox PARC’s pioneering word processor Bravo, word processors of the late 1970s also frequently were based on modes, such as create, edit, or insert, with completely different behaviors that may not be immediately obvious to the user.

A common joke2 about modes in word processors from that era went as follows: the user mistakenly typed EDIT while in command mode. Instead of the word “Edit” appearing, ‘e’ selected everything, ‘d’ deleted everything, ‘i’ told the word processor to switch into Insert mode, and the user then discovered that their entire document had been replaced by the letter ‘t’. Since “undo” was not yet a common feature found in word processors…this poor hypothetical user had no way to revert the change.

Ashton wanted his word processor to be friendly to use, avoid the pitfalls of bad interface design, and be as feature rich as he could implement within his few months of free time before classes started back up in the fall of 1977. By the end of the summer, Ashton didn’t really have a fully working word processor, but he did have a 50 page design specification, a pretty solid grasp of what he wanted to do, and confidence that implementing his design requirements would result in a markedly superior word processor. At this point summer break was over, so he called his largely theoretical program “WP” and went back to teaching classes.

But in spring of 1978 Ashton got a call from a man by the name of Don Owens, who asked Ashton if he would consider writing a word processor for the software company he was trying to get funding for to get off the ground. Don had chosen the incredibly 70s name of Satellite Systems, Inc for the company, apparently largely because he had seen the acronym “SSI” on the side of a boxcar and liked the way it looked.3

This word processor would be solely targeted at Data General minicomputers. Owens had heard of Ashton from the consulting work that Ashton had previously done for the air force and was undoubtedly happy to discover that Ashton had already put a considerable amount of thought into developing an excellent word processor.

Based on the promise of funding from Owens and a partner, Ashton hired his graduate assistant Bruce Bastian to work with him to turn Ashton’s 50 pages of project specifications into a fully fledged commercial product.

Unfortunately, right after Ashton and Bastian had signed a lease for office space, ordered a minicomputer for development, and Bastian had taken out a mortgage for a new home…Owens’ partner was unable to come up with his share of the funding and the deal fell through. Ashton had his work as a teacher to fall back on, but Bastian now had no job, and so Ashton reached out to his business connections to try to find a job for Bastian. Before too long he found out that a company called Eyering Research was looking for a new programmer.

Thanks to Ashton, Bastian was offered the job, and he soon discovered that Eyring had taken on a contract to provide a Data General computer to Orem, Utah with a new word processor. A word processor that they very definitely did not have and were going to struggle to provide within the contracted time.

Seeing an opportunity, Ashton and Bastian made a proposal to Eyring. Eyring would pay Bastian’s salary while he and Ashton worked on the word processor, which they would deliver to Eyering without further charge. Eyering would get the use of Ashton’s considerable expertise for free and would have the right to sell the resulting word processor as part of its other Data General sales, but Ashton and Bastian would retain full ownership of the product.

Eyering jumped at the opportunity and Ashton and Bastian set to work, with Ashton handling most of the printing code, Bastian developing the display code,4 and both of them working on the underlying code that tied it all together. The end product consisted of both men’s hard work and design decisions, and was delivered to Orem on time. The initial version of WP met Orem’s requirements nicely, however Ashton and Bastian felt that there was a lot more that they could add to the program.

In exchange for a free upgrade once it was completed, the city of Orem allowed them to continue using their Data General computer for free for another year, during which time Bastian borrowed money from his dad to keep afloat and Ashton kept his day job as a teacher and worked on WP on nights, weekends, and holidays.

To bring in some money while working on the full commercial version of WP, Ashton and Bastian decided to release a more basic version of WP aimed at programmers who just needed a good text editor, this version of WP to be called P-Edit. Since neither of them had any experience selling software, they reached back out to Don Owens, who was interested in selling both P-Edit as well as the full version of WP once it was ready. So the three of them incorporated in September of 1979, using the same SSI initials that Owens had previously chosen, but this time it was called Satellite Software International instead of Satellite Systems Inc.

By March of 1980 SSI released the full version of WP, now called SSI*WP, aimed at the wider commercial market. Or at least the wider commercial market of Data General software. The “WP” still didn’t really stand for anything apart from probably “Word Processor.”

The product gained good word of mouth and the three partners realized that they had something worth selling more generally on their hands. Priced competitively for that era and type of minicomputer software, they sold several copies a month at 5500 dollars apiece. Within another year, SSI had reached annual sales of 850,000 dollars.5 Nothing too earthshaking, but solid sales numbers especially when you consider the fact that it only cost SSI 25 dollars6 to produce each boxed copy of SSI*WP.

SSI*WP ran exclusively on Data General minicomputers, specifically ones running the AOS operating system, as opposed to other competing operating systems for Data General computers such as RDOS. It also required using a Data General terminal and printing could only be done to a Diablo 1650 printer or a clone of it.

SSI also eventually wound up with a rather odd management structure, with Ashton and Bastian each having 49.9 percent of the company, and the ostensible CEO, Pete Peterson serving as tiebreaker and possessing 0.2 percent of the company while Don Owens was pushed out.

Who was Pete Peterson, where did he come in, and what happened to Don Owens you ask? Let’s go back a couple years. In 1980 Peterson had been in a rough spot financially speaking. Together with a couple family members, he owned a custom drapery store and he spent his week going to appointments with housewives and trying to sell them on the virtues of paying for custom draperies. On weekends Peterson handled keeping the company’s books and taking care of the other paperwork such as creating the work orders for the sales he had generated during the past week.

It was a hard job at the best of times, and the 1980 recession had hit the little business hard, with sales collapsing by two-thirds and Peterson and his partners doing without pay for weeks as they struggled to keep the business afloat. He wound up having to take a part time job at a local grocery store to pay for his mortgage, where he earned a princely four dollars an hour.

But Peterson was far from an interior designer by training, he was actually a 1972 graduate of BYU with a degree in psychology. Prior to working at the family drapery business, Peterson had spent a year working as a bookkeeper before returning to his home state of Utah. His goal was to work work in the family drapery store (Julie’s Draperies) in order to save up the money to return to BYU in pursuit of a PhD. His plans changed a bit when he got married and started a family, and he then decided to focus on making a living for the time being, and put his plans for further education on hold.

Unfortunately in 1979 Peterson’s mother passed away, leaving Julie’s Draperies split into thirds between Peterson, and his brother and sister.The siblings tried to continue the business but none of them had the talent for creating new designs and marketing them that their mother had possessed and the business began to struggle.

Almost Perfect, Peterson, W.E. Pete, eBook - Amazon.com

As Peterson describes in his interesting little book (AlmostPerfect: How A Bunch of Regular Guys Built WordPerfect Corporation), it seemed like a hopeless situation: “I do not think any of us enjoyed owning or running a drapery business, but we were resigned to our fate.”7 The 1980 recession then gave the struggling business a further hefty kick and the business began to seem untenable, especially with trying to support three families.

This was when Peterson’s brother-in-law entered the picture - Bruce Bastian. As previously mentioned, SSI was doing well for itself in 1980, however although sales were solid, Bastian and Ashton were struggling to build the company correctly, and Bastian in particular was running himself ragged trying to do a bunch of office and HR work during the day and then code more on SSI*WP at night.

Bastian and his family frequently were over at Peterson’s for dinner and during one dinner, as Bastian described his workload, Peterson suggested hiring an office manager to relieve Bastian of his burden of office work, allowing him to focus on programming. Several days later, Bastian called Peterson and suggested that he take the job himself, part time but at a dollar an hour more than Peterson was making at the grocery store he had begun working for in order to help make ends meet..

Peterson was the sixth person hired at SSI, but was the first to quit. His first day at SSI he discovered that SSI did not do tax withholdings for its employees, figuring that employees could just sort things out on their own with the IRS. Since Peterson possessed an extremely healthy paranoia for getting on the wrong side of the IRS, he refused to get tangled up with the inevitable consequences of such an action.

When the three owners of SSI asked Peterson his reasons for immediately quitting, he gave them an on-the-spot education in employees versus contractors, bookkeeping, and basic business accounting.8 That same day they offered Peterson the full time salaried position of financial manager, with the authority to straighten everything out and make sure SSI was run properly.

And this example right here shows how valuable a man Peterson was to SSI. His experience as a small business owner had not only taught him the value of watching every penny, but had also taught him the necessary fundamentals of running a small business from top to bottom, from bookkeeping, to payroll, to organizational documents, expense categorization and even the necessity of stock certificates for the owners. Peterson even had to get SSI a business license as it was operating without one. His experience would prove invaluable in keeping SSI afloat and handling the rapid growth that the future held.

But what about Don Owens? To make a long story short, Owens’ vision for SSI began to widely diverge from Ashton and Bastian’s. It's important to make it very clear though, that Owens had been an invaluable part of getting SSI up and running, securing funding, selling P-Edit and SSI*WP copies, and doing the hard legwork and cold calling necessary to make SSI successful. Without his sales expertise and hard work, it's very unlikely that SSI would have been successful, just as without Ashton and Bastian’s hard work, SSI wouldn’t have had a product to sell or reason to exist in the first place. Owen also not only voluntarily stepped down from the board of directors in April of 1982, but also nominated Pete Peterson to take his spot.

The events that led up to this however were the differences between Owens on one side and Ashton and Bastian on the other. Disagreements over how Owens’s wanted to spend SSI’s money, how he kept overselling the product to potential customers by claiming that features that either did not exist or were still under development were actually usable, and the biggest of all, Owens’ desire to do a final round of financing and sell off a significant amount of stock. This would dilute the three men’s ownership of SSI, but would give SSI a lot more money in the bank and give each of the three owners a significant cash out.

This last one was the biggest bone of contention, and it caused even more issues as Owens began spending less and less time at the office and more and more time traveling around looking for investors, even though Ashton and Bastian repeatedly stated that they were not interested in taking on investors at this time or diluting their ownership. As friction between the partners continued to grow, matters were not helped by Owen starting up a new side business selling computer terminals and printers to the federal government. This inevitably led to concerns that SSI was paying for Owen to travel around selling not only WP but also his own side business hardware.

To make a long story short, eventually Owens was first removed from management, and then his ownership of SSI was diluted from one third all the way down to five percent before being bought out for 139,000 dollars in January 1983. When the inevitable court battles were over, Ashton and Bastian were left in full control of the company, with each owning 50 percent. Peterson was given the option to buy 0.2 percent and the trio then ran SSI for the next decade.

Although the sales of the still clunkily named SSI*WP were still doing well, Data General in particular and the minicomputer market in general was not where the big money was. Although the minicomputer market in the 1980s was healthy, the personal computer market was in the middle of experiencing exponential growth. Seeing the potential, and spurred on by requests from companies who were already using SSI*WP on their Data General minicomputers and wanted to be able to use it on the IBM PCs that they were starting to purchase, SSI started exploring the possibility of porting to the IBM PC.

In February of 1982 SSI purchased an IBM PC and set a newly hired BYU graduate, Alan Brown, to work exploring the possibilities of porting SSI*WP to it. Brown was a talented programmer, one of the many fruits of Ashton continuing to teach computer science at BYU as he could extend job offers to his best students.

Since the IBM PC lacked advanced programming tools at this time, Brown did a lot of the initial evaluation of porting SSI*WP by coding to the bare metal in machine language.9 Or at least this is what Peterson’s book AlmostPerfect states, although I suspect that Peterson (a non programmer) might be mixing up machine language with assembler. It would make sense for Brown to program the initial test of SSI*WP in assembler as that was what the full version was eventually programmed in. However, given the fact that an assembler for MS-DOS wasn’t available at this time, it's also possible that Brown was using a minicomputer development system that emulated the 8080 instruction set, I am honestly not sure. Or he may actually have programmed it all in machine language, which would be even more impressive.

The conclusion was that it was very doable to port SSI*WP to the PC, however in order to make it as fast and responsive as possible, it was decided not to program it in a high level language like C, but rather do it fully in assembler.10 This would give significant speed advantages and still be much faster than trying to do everything in machine language directly, but also forced SSI to wait until May of 1982 to really start work, as that was when the first assembler came to market for MS-DOS.

While work on the IBM port continued, SSI’s other programmers were working on the version of SSI*WP for the existing Data General market, by now up to version 2. This second version had a host of improvements including a spelling checker, the ability to add footnotes, and a very basic way of doing a very basic embedded spreadsheet with some arithmetic features for working with columns and rows.11 This was definitely not designed or marketed as a competitor to VisiCalc, it just was a handy little feature to have in a word processor.

In November 1982 WordPerfect 2.20 was finished and ready to be released for IBM computers. It was feature identical to the Data General version, but changed some of its commands to more friendly terms, such as replacing abort and kill with stop and cancel.12 By this point there were a number of competing word processors for the IBM PC on the market, ranging from cheap ones like Volkswriter to the market leader, Micropro’s WordStar which had about 75 percent13 of the PC market.

Before release though, a new name had to be decided on as while SSI*WP worked well enough for the comparatively small Data General minicomputer market, a shorter and more memorable name was needed for the mass market. Peterson suggested “WordPerfect”, however nobody else agreed with him and in a company wide vote, it came in close to the bottom, with the list being headed by Word Plus and ProWrite.

These two were to be sent to an attorney for trademark search, and Peterson included WordPerfect on the list as well, just in case. It turned out that both Word Plus and ProWrite were existing trademarks, so WordPerfect won by default, although Peterson was the only one enthused by this.

With the release of WordPerfect for IBM PCs and compatibles, sales jumped up strongly, although WordPerfect was still only a modest success as of yet. Coming out in the final quarter of 1982, it boosted that quarter’s revenue from the previous 200,000 dollars to 450,000 dollars.

October 1983 brought the release of WordPerfect 3.0 for IBM computers and the compatibles that were rapidly springing up. WordPerfect was receiving strong reviews and praise from not only tech publications but also from its users, who loved the fact that SSI had set up a toll free telephone line for support. The fact that WordPerfect also supported several hundred printers14 was a huge selling point in its own right as well. Keep in mind that in the early 1980s, applications were usually responsible for handling print drivers by themselves, MS-DOS did not do it for them. So printing to several hundred printers was the result of SSI painstakingly building and testing the drivers for every one of these printers all by themselves.

Even better was the fact WordPerfect 3.0 did a whole lot more to improve its printer compatibility than simply make it possible to print to more printers. All versions of WordPerfect 2.0 treated any supported printer apart from Diablo and Epsons as a dumb printer. This is not an issue that affects printers today, but once again, it was a common part of personal computing of that era. A dumb printer was one that only printed standard characters at a fixed width, no font choices, proportional characters, or bold or italicized text allowed. Even if the printer itself was capable of that, WordPerfect 2.0 didn’t make use of it. However in WordPerfect 3.0, supported printers could make use of all of these features, although a side effect of doing this meant that WordPerfect 3.0 was noticeably slower to print than previous versions.15

SSI also did a few ports of WordPerfect to the other personal computers of the era, although the target platforms picked were rather odd. The Tandy 2000 for example, which was marketed as an IBM compatible that ran MS-DOS and indeed it was compatible with all MS-DOS programs apart from the ones that didn’t work. Which was about half of them…best case scenario. Hence why the Tandy 2000 failed. WordPerfect was also ported to the Victor 9000, the DEC Rainbow, and the TI Professional, none of which went anywhere.

Thanks to the growing IBM and IBM compatible market however, WordPerfect was steadily growing in market share. The problem was that SSI was still primarily selling to dealers, not distributors, meaning that their distribution remained somewhat limited. SSI had tried to interest distributors in carrying WordPerfect, but had been turned down by every one they had spoken to, including large ones like Softsel. Things started to change when the large retail chain ComputerLand started stocking WordPerfect, which was not only helpful for moving more copies, but also gave WordPerfect a significant reputation boost as ComputerLand was a significant part of IBM’s dealer network.

SSI also cleverly gave every ComputerLand employee the chance to buy WordPerfect for only ten dollars, which helped incentivize them to try it, see its superior feature set to WordStar, which was what most of them were using, and thus get them to start recommending it over WordStar to customers. Within a few months, WordPerfect was ComputerLand’s second best selling word processor16, although WordStar was still in the lead. SSI closed out 1983 with 3.5 million dollars in sales, more than triple its sales for 1982.

Growth continued to rapidly accelerate as 1984 began, with sales rising to over two million dollars during the first quarter. In his book, Peterson describes the company at this time period as feeling like an out of control train as the company frantically struggled to keep up with demand. Still extremely cautious and wary of overspending, Peterson kept spending low and maintained SSI’s cash reserves as much as possible, driving the company to keep expenses down to a million dollars for the first quarter, just in case the massive jump was an aberration and the extra million was needed to keep the company going. SSI was still privately held by Ashton and Bastian, plus Peterson’s 0.2 percent share, and thus had the double edged sword of the freedom to run things as the trio saw fit, paired with a lack of the capital that would have come from a successful IPO or venture funding.

November 1984 saw the release of WordPerfect 4.0 for MS-DOS compatibles, and this is really when WordPerfect began to explode in popularity. At the start of 1984 they had only had about one percent17 of the personal computer word processing market, a fairly fragmented market that was dominated by WordStar’s 23 percent market share. However WordStar was beginning its long collapse, helped along by a baffling decision to cannibalize their own market share with WordStar 2000. By 1986 WordPerfect had grabbed 30 percent of the personal computer word processing market, and was still rising. Given WordPerfect’s explosive success, Satellite Software International made the sensible choice to rename itself to WordPerfect Corporation that same year.18

In May 1988 WordPerfect 5.0 was released for IBM computers and compatibles. Sticking to their guns, WordPerfect 5.0 was still exclusively targeted at MS-DOS computers, and a Windows version was still not even under consideration. Of course at this point Windows was only at version 2.1, and was not selling in large numbers, with all versions of Windows totaling fewer than two million sales across all available versions.19 The MS-DOS application market was hot, and WordPerfect was riding high, with no real competitors threatening them.

WordPerfect 5 got off to a very rocky start since it was not only released with a number of bugs, but one of the bugs was in the installation process itself. Essentially two unneeded files were removed from the WordPerfect install disks at the last minute in order to save space and prevent an additional disk from being needed in every box. These two files were completely inconsequential and their absence didn’t cause WordPerfect itself trouble…however the 5.0 installer was never updated to account for these two missing files.

WordPerfect would still install correctly, but it would throw two error messages during install that looked ominous. WordPerfect 5.0 was released to distributors on Thursday, May 5th, 1988. The calls began on Monday, May 9th and reached a crescendo on Friday, May 13th. At noon, with all 1-800 lines into Utah busy, AT&T called WordPerfect’s headquarters to, in the words of Pete Peterson “politely inquire about when we thought we could clear up our busy signals.” 20

Specifically, AT&T’s full 1-800 line Utah capacity was vastly insufficient to meet the load, which was so great that both Delta Airlines (reservation center) and American Express (customer service center) were essentially cut off from the outside world. AT&T eventually wound up adding quite a bit of additional 1-800 line capacity to Utah as a result of this event.21 Some sources seem to confuse the overload on the 1-800 lines with a failure of the Utah phone system as a whole, but as near as I can tell, normal phone lines were not affected.

Still, even with all the bugs, WordPerfect 5 was a huge success, with WordPerfect itself seeing increased revenues of 75 percent over the previous year.22 There seemed to be no competitors able to touch it, and the decision to avoid wasting development resources on a Windows port seemed validated.

Sure Microsoft Word for Windows was in development, but it wouldn’t be released until 1989, and the MS-DOS version of Microsoft Word had posed little threat to WordPerfect’s increasing domination of the market, regularly appearing below WordPerfect in comparisons from tech publications. The fact that Microsoft Word was considered the best word processor for the Macintosh, and (more importantly) that Microsoft now had years of experience in developing a word processor that ran under a full GUI, was not considered interesting or relevant to WordPerfect. They had their hands full with a rapidly expanding market that was eagerly buying more and more copies of WordPerfect, competitors that were falling further and further behind, and (apparently) good news everywhere they looked, with nary a cloud in the sky.

There were also many other versions of WordPerfect released throughout the 1980s, with a rather confusing numbering scheme that did not track with the main DOS releases. These other releases also wasted valuable time and effort for a very small reward. For example, 1988 alone saw ports of WordPerfect released for the latest Data General minicomputers, DEC, the Macintosh, and AT&T’s Unix. Apart from the Macintosh version, none of these were really worth the effort, although since WordPerfect had started on Data General computers, I’m sure their continued support of the Data General platform was appreciated.

WordPerfect’s attitude towards Macintosh users was to call them “Macintosh bigots”23 (Pete Peterson himself used this term) who for some reason didn’t appreciate WordPerfect’s lackluster attempt at a Macintosh program. This attempt basically saw the clumsy DOS interface ported straight to the Mac rather than any attempt made to produce a new interface in line with Macintosh design principles.

Additionally, WordPerfect for the Mac used the older WordPerfect 4.1 format for its files, rather than the newest 5.0 format (WordPerfect 5.0 was released for DOS one month later), causing issues with file sharing between DOS and Mac users. Upon its initial release in April of 1988, Macintosh WordPerfect sold decently, but not well, eventually reaching about ten percent of the Macintosh word processing market.

The Macintosh version of WordPerfect could have been a useful training ground for turning a complicated DOS based application into a smooth GUI driven experience, however WordPerfect never grasped this, leaving Microsoft to reap the benefits down the road as it took the lessons it had learned with the mature Macintosh version of Microsoft Word, and ruthlessly applied them to the Windows version of Word.

However WordPerfect Corporation did grudgingly understand that they should probably do a GUI version of WordPerfect at some point…but they struggled to decide whether to do a new version of WordPerfect for Windows or OS/2. Eventually they wound up taking the solomonic approach of splitting WordPerfect’s development efforts and doing a version for both.

The first public result of this bifurcated effort was a very buggy, unfinished version of WordPerfect 5.1 for OS/2 that was demonstrated by Ashton at COMDEX 1989. Ashton vaguely committed to a promised release date of 1990 for the OS/2 version, and a Windows port was presumably sort of hanging out there in the ether somewhere.

According to Barbarians Led By Bill Gates, WordPerfect (and Lotus’s) attitude towards Windows was that it was an “...unproven platform receiving faint endorsement from its own maker while facing imminent death at the hands of IBM's OS/2.”24

It seems that the Windows version of WordPerfect was still at the bottom of the priority list, with very little effort and no focus or urgency lavished on it. OS/2 was where WordPerfect’s lackluster GUI development efforts were being spent, and this was about to bite them in the butt, hard. In January of 1990 WordPerfect had given into Microsoft’s offer to be a beta site for Windows 3.0, which could have been a significant opportunity. The chance was there to push hard, and get WordPerfect for Windows written and out the door in time for Windows 3.0’s release, when it would benefit from the burst of publicity and marketing that Microsoft was pushing for what promised to be the most significant release of Windows to date.

Microsoft was also pushing for as many high end MS-DOS applications as possible to be made available for Windows 3.0, knowing that this would help drive adoption, and would have been happy to help push WordPerfect, had it been available at Windows 3.0’s release. Getting the top MS-DOS word processor on Windows would have been a major marketing coup, and one that WordPerfect would have greatly benefited from. Microsoft Word did exist at this time, but it was definitely an also-ran with a fraction of WordPerfect’s market share, and Microsoft was more eager to drive Windows adoption than they were to drive Word adoption.

Unfortunately this priceless opportunity was missed and it was missed for two painful reasons that resulted in WordPerfect’s dominant marketplace position being effectively destroyed in a very short period of time. First of all there were precisely zero experienced Windows developers at WordPerfect…zero. It simply hadn’t been a priority to recruit any and no attention had been paid to getting anybody internally up to speed.

This meant that the OS/2 development team was the one tasked with doing the Windows port. And this team, for all the talented and capable people in it, was not WordPerfect’s A team, as the A team wanted to be working on the DOS version of WordPerfect, which was where most of the prestige still was. Creating an application for either OS/2 or Windows meant a lot of hard work, a lot of dead ends, and the end result would be an application that ran on only a minority of computers, instead of the massive install base of MS-DOS machines where WordPerfect had already crushed the competition.

And developing applications to run correctly under early versions of Windows was extremely difficult, partially due to Windows own idiosyncrasies, some of which are still hanging around to torment developers even today. But even apart from the difficulty of dealing with Windows as a developer, and Windows as a GUI, was the difficulty of dealing with the paradigm shift that Windows represented. DOS programmers were used to being in basically full control of the PC, down to the hardware level. MS-DOS was a very VERY basic operating system, and one that gave the programmer enormous flexibility to work with the hardware (for better or worse). Under Windows, much of that flexibility was gone as it interposed itself between the application and the underlying hardware.

Windows 3.0 didn’t even let programs go full screen, one of many reasons why games for Windows didn’t really become a thing for a number of years. Working with the new Windows APIs, working with a full WIMP interface, retooling a DOS application’s interface to take full advantage of Windows interface paradigm, and accepting the inevitable significant performance hit that an application running under Windows versus running under DOS suffered, required a great deal of work to learn and many software developers simply weren’t interested.

In his entertaining book In Search of Stupidity, Merrill Chapman cites an interview with Joel Spolsky, a member of the Excel development team in the early 90s and the founder of every programmers’ best resource, StackOverflow. In this interview, Spolsky stated that the WordPerfect for Windows development team also hobbled themselves by continuing to program in assembler, rather than a higher level language like Pascal.25 This was one of the major reasons why the MS-DOS version of WordPerfect was so responsive on the slow processors of the 1980s, but added significant development time on top of the WordPerfect team’s existing struggles with programming for Windows.

Granted, working with Microsoft was always a risky proposition, something that Pete Peterson was very aware of, having compared Microsoft to “a fox that takes you across the river and then eats you”26 years prior, and he was also undoubtedly aware of Bill Gates’ practice of twice a year going to WordPerfect’s largest customers and personally demanding to know why they hadn’t switched to Microsoft Word yet. There was a lot of justified wariness about working with such ruthless competitors, and it was easier to just ignore Windows and focus on MS-DOS.

There are also rumors that the 1991 expanded FTC investigation of Microsoft was spurred on by complaints from WordPerfect to the FTC about Microsoft’s anti-competitive behavior.27 Such rumors undoubtedly also reached Gates’ ears, and probably did nothing to mellow his attitude towards crushing WordPerfect as soon as an opportunity presented itself. Not that Gates ever needed any specific reason to do his utmost to crush competition.

WordPerfect 5.1 for Windows finally staggered out the door towards the end of 1991, in a rather lackluster condition. It was slow, it was buggy, it was apparently missing features, it had to be installed from MS-DOS rather than within Windows, and worst of all it apparently had significant stability issues. This led to a lukewarm reception from consumers and possibly reviewers, although I was unable to find anything in InfoWorld or Byte actually slamming it. Wikipedia does state that it was “largely unpopular”,28 but provides no links to sources for this, one of many reasons why I do not rely on Wikipedia for research.

Getting decent initial reviews for a lackluster product was possibly due to WordPerfect coasting on its existing reputation as the leading MS-DOS word processor. InfoWorld’s review even gave it the rather left-handed compliment by saying that while it was “an excellent value” it was also “less powerful than Word for Windows 2.0”,29 while Byte magazine enthusiastically called it “a solid product backed by a solid company; it's particularly attractive if you use other computer systems running WordPerfect. Highly recommended.”30

InfoWorld’s review stating that Microsoft Word was superior to WordPerfect for Windows was a harbinger of the giant collapse that WordPerfect was about to experience. Whether WordPerfect for Windows was favorably reviewed or not, Microsoft Word had finally closed the feature and perception gap and was seen as the better product of the two. And consumers typically tend to avoid buying a product (all other things being more-or-less equal) that is seen as second best.

WordPerfect was heading into a crisis, and stern measures would need to be taken if they were to overcome Microsoft Word and maintain their edge as the superior product. The MS-DOS market was falling, and the Windows market was rising, with more and more people starting to adopt it, especially since the first genuine Windows juggernaut, Windows 3.1, had been released in March of 1992.

What WordPerfect did have going for it was a solid reputation, extensive experience in developing word processors, and more importantly a giant pile of cash totaling more than 100 million dollars. Thanks to Peterson’s careful management and cost controls, WordPerfect’s substantial war chest had positioned the company to come back from their poor start on Windows. With deep pockets, an efficient company with comparatively low overhead, extensive experience with developing word processors, a business that was still growing and making money, plus the right strategic direction, WordPerfect still had a fighting chance of coming back and reclaiming their crown.

Then Bastian and Ashton decided to shoot themselves and WordPerfect in the foot and get rid of Peterson. In the words of In Search of Stupidity, they were “tired of Peterson’s stern-daddy management style [and] decided they were ready to fly free on their own.”31 According to Accidental Empires, there was also a significant disagreement, as Basian and Ashton wanted to primarily focus on out-marketing Microsoft, whereas Peterson wanted to focus on product development instead, and make sure WordPerfect regained its competitive edge over Microsoft Word before spending vast sums of money advertising.32

So Peterson was sent packing in 1992, and with his departure, WordPerfect lost the man who best understood the necessity of keeping costs as low as practical, keeping employee numbers reasonable, and making sure that the business was run without any excess fat. Ashton and Bastian promptly went on a major hiring spree, increasing the company’s size by a mildly shocking FORTY percent from 1992 to 1993. Although revenue increased at the same time, still buoyed by WordPerfect’s domination of the still reasonably healthy MS-DOS word processing market, sales growth slowed and the increase in revenue was more than canceled out by the massive increase in expenses as the WordPerfect Corporation began to bloat.

Microsoft then gave the company a hefty kick in the face in 1991 by releasing the first version of Microsoft Office, which combined, among other things, the very well reviewed Microsoft Word, plus the leading spreadsheet, Excel, plus PowerPoint, all bundled together for 495 dollars. This overwhelming attack on multiple fronts immediately started to do serious damage to Microsoft’s application competitors, damage that was only increased when Microsoft added its new database, Access to Office in November 1992.

Keep in mind that whereas Office gave you three powerful programs for 750 dollars, the Windows version of WordPerfect launched at 495 dollars and that only gave you…WordPerfect. Given a choice between spending almost 500 dollars on a single program, or half again as much for three programs, at least two of which were the best in their class (I’m not exactly sure where PowerPoint ranked against the competition in 1991)...well that’s the sort of pricing that tends to destroy the weaker product.

Again, at this point Microsoft Word was already at version 2, and as previously mentioned was reviewing better than WordPerfect for Windows, which was only in its first version. Microsoft’s extensive experience in developing GUI applications (all thanks to its work as an early, and leading, Macintosh developer) began to pay enormous dividends as Windows started to take off, and Microsoft was ruthlessly using it against its competition.

WordPerfect did have the mild consolation that it wasn’t alone in flailing wildly against the Office juggernaut, but this didn’t exactly translate to improving its collapsing market share.

Since WordPerfect still only possessed a word processor, Ashton and Bastian struck an agreement with another company that Microsoft was demolishing with the Office suite, Borland. The agreement saw Borland’s spreadsheet Quattro, plus their database Paradox, mashed together with WordPerfect to create a sort of Frankenstein's monster office suite that was called “Borland Office” and launched at a price of 595 dollars in 1993.33

The market took one look at this and laughed hysterically all the way to fattening Bill Gates’ bank account further. The painful fact was that cobbling an Office competitor together from products that were already losing separately to Microsoft’s applications…merely ensured that they lost collectively. Additionally, the applications that comprised Office were designed to work together to a greater or lesser extent, and had a fair amount of cohesion between the products.

By contrast, WordPerfect and Borland’s ad-hoc office suite consisted of products that were never designed to work together, had noticeably varying interface conventions, and did not meaningfully share data between applications to the extent that Office did. For example, Office allowed the user to embed (using OLE) a spreadsheet from Excel into a document in Word, which would automatically update in Word if the spreadsheet was changed in Excel. Quattro and WordPerfect did not really share this ability at all, so far as I can discern.

Now, in fairness I must add that Borland and WordPerfect didn’t simply print some new software labels for the diskettes, create a spiffy box for the applications to come in, and make some new logos before calling it a day.

There was an effort made to make the different applications work together, through a “middleware” software solution called PerfectFit that WordPerfect wrote. However it seems that PerfectFit was not part of the initial 1.0 release of Borland Office, but rather was only included as part of the 2.0 release in 1994. Lacking this meant that the initial release of Borland Office was severely lacking in integration, just when it needed to make the best initial impression.

Still, PerfectFit’s inclusion in the second release of Borland Office was called out favorably by ComputerWorld in a review, stating that PerfectFit was “an upgrade to the Borland suite featuring tighter integration between Borland applications and WordPerfect Corp.’s namesake word processing software…”,34 although I do find it significant that they said “tighter” integration instead of “tight” integration.

PerfectFit did its level best to tie the varying applications together, but it really had a thankless task and so far as I can tell never approached the level of integration that Microsoft Office’s bundled applications enjoyed.

And while some of this was probably (certainly) due to Microsoft’s tendency to use undocumented features and API’s of Windows in order to give their products a further competitive edge, the fact remained that to the end user, Microsoft Office’s included applications worked well together, and Borland Office’s bundled applications didn’t.

Novell then bought the flailing WordPerfect Corporation in 1994 for 745 million dollars,35 along with purchasing Quatro Pro from Borland, from whom they also licensed relational database Paradox. Novell’s plan was to rebrand Borland Office and try to successfully take on Microsoft Office with a...name change.

This time around the former Borland Office suite would be known as WordPerfect Office or just PerfectOffice. I’m assuming that the thinking behind this was that “WordPerfect” had better name recognition than “Borland”, and thus would make marketing easier.

The scope of what Novell was trying to do to integrate the various applications into a coherent whole that also played well with Novell’s existing applications was so ridiculously complex and has so many pieces that it really is no surprise that it failed.

I’ll just hit a couple of the high points here, drawing heavily from an article in Computerworld with the incredibly gripping title of WordPerfect, Novell consolidate groupware efforts. This article says that Novell’s key strategy was that the components will plug into other vendors’ applications and services. One example given was that WordPerfect Office will plug into Novell’s directory service, and that the WordPerfect Office’s email functionality would be merged with Novell’s similar product (Message Handling Service) to form a new singular product that would be called the Open Messaging Environment.

Novell’s plans to make WordPerfect Office and its other applications into a sort of plug-in architecture that could interface with any number of third party applications in some sort of vaguely specified way was not exactly the right mindset for success against Microsoft's laser-like focus on savaging the competition.

Hence Novell quickly wound up throwing in the towel on WordPerfect Office after a mere two years, and sold it on to Canadian based Corel for 124 million in 1996, meaning that they had either waaaaay overpaid for it in 1994 (this is probably correct) or it had lost five-sixths of its value in only two years (may need to embrace the power of ‘and’ on this one).

WordPerfect’s market share was in freefall during this time, dropping from a still healthy fifty percent market share in 1995 all the way down to a mere five percent by 2000. The depth of WordPerfect’s collapse is also laid out in a 2000 New York Times article, nicely encapsulated in the quote “Today many people are surprised to learn that WordPerfect still exists, so complete is Microsoft Word's dominance.”36

In only five years Wordperfect had gone from significant market dominance to “pursuing a strong niche strategy.”37 This quote came from Dave Ludwick (one of Corel’s senior managers), in an interview for the aforementioned NYT article, where he made a heroic effort to make lemonade out of lemons. But his attitude accurately represented Corel’s determination to valiantly keep going on, trusting in its hard core of dedicated fans to keep WordPerfect a viable business. Corel would surprise everybody by finding enough success with this approach to remain in business to this very day.

That’s right, you can still buy WordPerfect today, and not just as a digital download either, but actually as boxed retail software, available at places such as OfficeDepot and Best Buy. WordPerfect itself can be bought for 49.99, or as part of the Corel office suite, called WordPerfect Office 2021. This office suite comes in a couple flavors, and is built around an updated version of what was originally the (of course) long-defunct Borland Office.

My local Best Buy even stocks WordPerfect!

It is now a nicely integrated set of applications however, with consistency across all of them, enabling it to do a good job of competing directly with Microsoft Office on a feature-by-feature basis. Similar to Microsoft Office, WordPerfect Office is available in both Home and Student as well as Professional editions.

There is WordPerfect for word processing of course but there is also an update of the venerable Quattro Pro for spreadsheet work. Quattro was an original part of the Borland Office suite, and it's nice to see it still around and kicking. There is also the equally venerable application Presentations for, well, making presentations a la Powerpoint, and a couple other applications that have no direct counterpart in Microsoft Office, such as the lightweight photo editor AfterShot 3.

The modern WordPerfect Office is solely available for Windows however, as the last major Macintosh version (3.5) was released in 1996, with a minor update in 1997. It was never updated for Mac OS X and thus cannot be run by any modern Mac (Intel or M series) without resorting to a classic Mac (pre-Mac OS X) emulator such as Basilisk or Sheepsaver.

I’m not really sure what WordPerfect’s market share is these days, as it's hard to tell exactly how many people are using it solely, using it in conjunction with something like Google Docs, or maybe using it in place of Word while still using other applications from Microsoft Office. And of course there might be a couple people stubbornly clinging onto Quattro Pro and buying the full WordPerfect Suite just to get it.

A Google search for “wordperfect office marketshare” returns such diverse results as one from Slintel.com stating that WordPerfect Office has a 1.50 percent market share, which is contrasted with the results from Datanyze, which says WordPerfect is the number five ranked office suite(?) with a market share of 0.02 percent.

What can be said in confidence is that WordPerfect has enough of an audience that it is still profitable for Corel to sell in both digital and box form, and has managed to outlast almost all of its competitors. It stands as one of the very few applications that ever went head-to-head with Microsoft and managed to stay alive. Even in its current fractional market share, it still survives and is under active development, still serving people almost fifty years after Alan Ashton programmed the first prototype.

I’d say that’s pretty neat.

Another Boring Article

Discussion about this post