In 1984, a piece of software called WordStar controlled sixty percent of the word processing market on personal computers. Its publisher, MicroPro International, had just gone public and was, by at least one measure, the largest software company on Earth: seventy-two million dollars in annual sales and more than 650,000 copies in circulation. Every law firm, newsroom, and university department that owned a personal computer had a copy on a floppy disc somewhere in a desk drawer. At least half of those copies were probably pirated, which only made the dominance more complete.

The program had a signature its users knew by feel before they knew it by name: four keys, Ctrl-E, Ctrl-S, Ctrl-D, and Ctrl-X, arranged in a diamond under the left hand, moving the cursor up, left, right, and down without fingers ever leaving the home row. They called it the WordStar diamond. If you learned to write on a personal computer in the early 1980s, your hands learned that diamond the way a pianist’s hands learn the keys. It was not a feature.

It was a reflex. By 1993, MicroPro no longer existed. WordStar’s market share had fallen to two percent. The company had been renamed, gutted, and sold to SoftKey, which collected dead software brands the way a pawn shop collects other people’s wedding rings.

Today, almost no one outside a small circle of computing historians and a handful of stubborn novelists can explain what WordStar was, why it mattered, or how a company that owned the entire word processing market managed to lose it all because it could not ship a single meaningful update in four years. To understand how WordStar happened, you have to understand two men who came from very different directions and collided inside a company most people have already forgotten. Seymour Rubenstein was born in 1934 and grew up in Brooklyn. As a teenager, he repaired televisions, which in the late 1940s meant crawling behind wooden console cabinets to replace vacuum tubes.

He did his military service, put himself through college at night while working as a technical writer, and in 1964 helped design a classified acoustic identification system for the United States Navy. He was not a programmer. He was a man who understood systems, markets, and which technologies were about to move from expensive to cheap. After a stint overseeing software development for programmable terminals, he relocated to the San Francisco Bay area and eventually became director of marketing at IMSAI, a small company in San Leandro, California that made the IMSAI 8080, one of the first personal computers sold to hobbyists and small businesses in the mid-1970s.

It was there that Rubenstein met the second half of the equation. Rob Barnaby was a programmer who wrote code in Intel 8080 assembly language, as close to speaking directly to a microprocessor as a human being can get without soldering wires to the chip itself. Rubenstein had been watching the market for standalone word processors: the big dedicated machines made by Wang Laboratories, IBM, and Xerox that cost $10,000 or more. A Wang word processor in 1978 was office furniture with a price tag only a law firm or corporate headquarters could absorb.

In 1977, Wang sold over 20,000 of them. The market was worth billions and growing, but it was locked behind a price wall that kept small businesses and individual writers on the outside, still using typewriters and retyping entire pages to fix a single paragraph. Meanwhile, personal computers had almost nothing for document creation. There were text editors, barebones programs that let you type characters and save them to disc, but no formatting, no pagination, no way to see what a printed page would look like.

Rubenstein looked at the gap and saw a fortune. What if a $495 piece of software could do what a $10,000 Wang machine did? In June of 1978, he left IMSAI with $8,500. No venture capital, no angel investor.

He incorporated MicroPro International in San Rafael, California, then went back to IMSAI and recruited the one person he knew could build what he had in mind. Barnaby said yes. Barnaby started by writing two smaller programs, a text editor called WordMaster and a sorting utility called SuperSort, to generate immediate revenue. Then Rubenstein handed him a specification based on a study of the big word processors from IBM, Xerox, Wang, and Lanier: formatting, pagination, mail merge, on-screen help, all brought down to a microcomputer running CP/M.

What happened next is one of those feats of individual engineering the software industry has mostly forgotten. Barnaby sat down at an IMSAI computer and began writing alone. No team, no code review. Over the next four months, he wrote approximately 137,000 lines of Intel 8080 assembly language, where a single misplaced byte could crash the entire machine.

MicroPro showed WordStar publicly for the first time at the West Coast Computer Faire in April 1979 and began selling it that June at $495 plus $40 for the manual. For that money, a small business or freelance writer could now produce formatted documents, merge address lists into form letters, and see page breaks on screen: capabilities that had previously required a machine costing twenty times as much. In its first eight months, WordStar sold 5,000 copies. Rubenstein later described the division of labor bluntly: Seymour was the marketing brains.

I was the technical brains. That sentence matters, because the man who figured out how to do it and did it was about to disappear, and the company would spend the rest of its life trying to replace what he took with him. WordStar was the first word processor for a microcomputer to offer true mail merge, the first to show on-screen pagination with a full line of dashes marking page breaks, the beginning of what the industry would later call what you see is what you get. But the manual was a disaster.

InfoWorld called it incredibly inadequate, and the inadequacy had an unintended consequence: third-party authors started writing their own WordStar guides, and those books sold by the hundreds of thousands, introducing WordStar to people who might never have heard of it. Piracy compounded the effect. Versions through 3. 3 had no copy protection, and users copied floppies for friends and entire offices.

The worse the manual, the more books got written about it; the easier it was to copy, the more people used it; the more people used it, the more essential it became. Barnaby did not get to enjoy this. By March of 1980, exhausted from writing 137,000 lines of code in four months, he left MicroPro. For all practical purposes, the man who had built the entire product was gone before the product reached its first anniversary.

The company kept selling his code. Sales climbed from $500,000 in 1979 to $1. 8 million in 1980, $5. 2 million in 1981, and $23 million in 1982.

The revenue kept doubling, and the code he left behind sat at the heart of the company like a cathedral built by a single architect who had left no blueprints. No one else at MicroPro fully understood it. Then came the deal that turned dominance into inevitability. In 1981, Adam Osborne released the Osborne 1, the first commercially successful portable computer, and bundled WordStar with every unit sold.

Kaypro did the same. WordStar was no longer just the best word processor on the market; it was the word processor that came in the box. By the fall of 1983, cumulative sales had passed 650,000 copies, more than double the second-place product. WordStar 3.

3, released in June of 1983, was the peak. This was the version that 650,000 people had, the version bundled with their Osborne or Kaypro, the version they had pirated and passed to friends, the version they had learned on and often refused to unlearn. The pivot that should have locked this in forever came in April 1982, when MicroPro released WordStar 3. 0 for MS-DOS.

The IBM PC had launched in August 1981, and the world was moving from CP/M to DOS. WordStar made the jump early enough that its dominance carried over intact. By fiscal year 1983, revenue hit $45 million. By fiscal year 1984, $72 million.

The company went public. By one industry measure, MicroPro International was the largest software company in the world. But the seeds of everything that followed were already in the ground. MicroPro had no corporate sales program until December 1983, a notable failure for a company selling business software.

The manual was still terrible, and customer service was worse. Rubenstein had been building toward something larger: an integrated office suite called StarBurst that would combine word processing, spreadsheet, database, and communications tools into a single product, a radical idea years before Microsoft Office existed. But he stepped down as chief executive in August 1983, replaced by E. Glenn Haney, a former Sperry Corporation vice president.

In January 1984, Rubenstein suffered a heart attack, and whatever influence he still had evaporated. The new leadership cancelled StarBurst. The market was reshaping itself around MicroPro, and the company was not paying attention. In 1980, the installed base of personal computers in American businesses was measured in the tens of thousands.

By 1984, it was measured in the millions, each needing software. But this was no longer a market of enthusiasts who learned software through patience and repetition. It was millions of office workers who expected a manual they could read and a support line they could call. And in January 1984, Apple released the Macintosh, a machine with a graphical user interface, a mouse, and a word processor that showed text on screen in the fonts it would print in.

The future was visual. WordStar was text. Serious competitors entered the market at almost the same moment. WordPerfect, created by two professors at Brigham Young University, released its first version in 1982.

Microsoft Word appeared in 1983. Both companies understood something MicroPro did not: in a market of millions, the product that ships updates fastest and treats its customers best will eventually win. WordPerfect built its legend on a toll-free support line staffed by people who actually answered the phone and walked users through their problems as long as it took. Microsoft built its legend on relentless iteration, a ruthless willingness to make each version better than the last.

MicroPro, meanwhile, had the bestselling word processor in the world and could not figure out how to write the next version of it. WordStar 3. 3 was the peak. And the gap that destroyed the company deserves plain statement: version 3.

3 was released in June 1983. Version 4. 0 did not arrive until 1987. Four years.

In those four years, WordPerfect released five major versions. Microsoft Word iterated relentlessly. The market moved. The users waited, and then they stopped waiting.

In October 1982, MicroPro laid off a programmer named Peter Mero during a workforce cut of nearly half. Mero took his knowledge of WordStar’s architecture, recruited other former employees, and founded a rival company called NewStar. In September 1983, NewStar released New Word, file compatible and command compatible with WordStar 3. 3, a clone with improvements: a built-in spell checker, support for laser printers, and features users had been asking MicroPro for and not receiving.

A fired employee was beating MicroPro with its own product, improved. In December 1984, MicroPro released WordStar 2000. The name suggested evolution. The product delivered confusion.

It was not an update to WordStar but an entirely different program with a new interface, new commands, and no compatibility with existing WordStar files. A user who had spent years building a library of documents could not open them. PC Magazine said it had all the charm of an elephant on motorized skates. It shipped with copy protection, alienating the user base that had grown up on freely copied floppies, and it divided MicroPro’s own employees into warring factions.

There is a strange footnote: WordStar 2000 introduced three keyboard shortcuts, Ctrl-B for bold, Ctrl-I for italic, and Ctrl-U for underline, which today are universal across every word processor and email client on Earth. The program failed. The shortcuts survived. MicroPro tried to update the original WordStar codebase internally.

Without Barnaby, the code had become what industry people called spaghetti, tangled and undocumented and resistant to change. In the end, MicroPro bought New Word. WordStar 4. 0, when it finally shipped in 1987, was built on the code of a clone made by a man the company had fired five years earlier.

The numbers tell the rest. In 1984, WordStar held sixty percent of the word processing market. By 1986, twenty-eight percent. By 1990, ten percent.

By 1993, two percent. WordPerfect rose from eleven percent to sixty percent over the same span. In 1989, MicroPro renamed itself WordStar International, as if changing the name on the building could change what was happening inside it. In 1991, it released WordStar for Windows, a competent product, but two years behind Microsoft Word for Windows, which had already established itself as the corporate standard.

Being good is not the same as being on time. In 1993, WordStar International merged with SoftKey. A company that once generated $72 million a year and held sixty percent of the market was absorbed into a rollup firm that specialized in buying distressed software brands at a discount. The name passed through a series of corporate acquisitions and reorganizations.

At some point, the WordStar archives were reportedly mislaid. No one quite knows where the original source code ended up. Seymour Rubenstein went on to build a spreadsheet program called Surpass, sold it to Borland, where it was renamed Quattro Pro, and founded two more companies in the 1990s. He died on November 12, 2020, at the age of 86.

But WordStar did not die with MicroPro. George R. R. Martin, author of A Song of Ice and Fire, told Conan O’Brien’s audience in 2014 that he writes on a DOS machine running WordStar 4.

0. Not a retro affectation, his working tool. Robert J. Sawyer, the Canadian science fiction writer who has won both the Hugo and the Nebula, wrote all 25 of his novels on WordStar 7.

0 for DOS. In August 2024, Sawyer packaged the complete WordStar 7. 0 archive and released it for free download, containing the software, the documentation, DOS emulators for modern Windows machines, and a command reference. The diamond lives on in places most users never notice: the cursor commands influenced editors like Borland’s Turbo Pascal, and dozens of modern text editors still offer WordStar-compatible key bindings.

The Library of Congress is now developing format description documents for WordStar files as part of its digital preservation initiative, working to ensure that files created in WordStar can still be read. Those four keys are still there. Ctrl-E, Ctrl-S, Ctrl-D, Ctrl-X. The diamond under the left hand.

WordStar did not fail because someone built a better word processor. WordStar failed because the company that made it could not do the one thing its own software was designed to do: finish writing something and ship it.