Hard Drive: Bill Gates and the Making of the Microsoft Empire (19 page)

Such was the state of affairs between Microsoft and Digital Research in September of 1980 when Gates picked up the phone and called Kildall. Once again, he was sending business to someone he now considered a competitor, but now he had no choice. IBM wanted CP/M as an operating system for its first personal computer.

What happened the day IBM came courting Digital Research in Pacific Grove, California, just off scenic Highway 1, has become part of the folklore of the personal computer industry, a story that is told over and over whenever the talk at an industry convention or dinner party turns to Microsoft and its money-making monopoly, DOS. Gary Kildall blew the opportunity of the century, someone will say. The tale is told how Kildall was out flying his twin-engine plane while the men in blue suits from IBM were waiting for him on the ground.

In the dozen years since that fateful day, many different versions have appeared in books, magazines, newspapers, and trade journals. Few of the accounts are exactly the same. The details usually differ in small ways and in large.

The two central players in the story, Gary Kildall and Jack Sams, remember the day’s events much differently. Sams said his group, which consisted of himself and two or three others from IBM, flew down from Seattle to talk with Kildall as planned the day after Gates made the call from his office. They were met by Kildall’s wife, Dorothy McEwen, who was actually running the business at the time while Gary handled the technical work. Also present was a lawyer from Digital Research. But Gary was not around.

“The meeting was a fiasco,” recalled Sams.

Before Sams could talk about why he was there, McEwen and the lawyer first had to agree to sign IBM’s standard nondisclosure agreement, the same one Gates had signed. They would not do so. To outsiders unaware of IBM’s policy with vendors, the agreement was somewhat intimidating, but it helped protect IBM from lawsuits. It stipulated that a vendor, in this case Digital, could not tell IBM anything confidential in the meeting, and if something confidential were revealed, IBM could not be sued if it acted on the information. On the other hand, if Digital revealed or acted on anything confidential it heard from IBM at the meeting, it could be sued.

“We tried to get past the point of signing this nondisclosure agreement so we could talk about what we came down to talk about,” said Sams. “It was three o’clock in the afternoon before they finally got around to the point of signing an agreement that [said] we had been there and they would not disclose it. I was completely frustrated. We went back to Seattle the next day and I told Bill we had been unsuccessful in trying to deal with Gary and would he see what he could do about getting a commitment from him for a 16-bit operating system.”

Sams said Gary Kildall never did show up before he and his group left Digital Research. Bill Sydnes, engineering manager for Project Chess, confirmed that Sams later told the task force Kildall would not meet with him.

An industry friend of Kildall’s, who did not want to be quoted by name because of their friendship, said he once talked with a member of Sams’ team who visited with Dorothy McEwen and Digital’s lawyer that day. “The IBM guy told me they had never seen anyone [McEwen and the lawyer] quite as rude and arrogant in their lives.”

Kildall acknowledged he was flying his plane at the time of the meeting, but he said he was on a business trip in the Bay Area. “The stories make it sound like I was doing loops or something. But I was out flying on business, just like someone else would be driving a car. I knew the IBM people were coming in.” Kildall claimed he was back at Digital by early afternoon, in plenty of time for the meeting with the IBM group.

“My wife had some concerns before I arrived, sure. If you sign this agreement, it says they can take any of your ideas and use them anyway they want. It’s pretty scary. My wife had never seen anything like that before. I explained that these were not bad guys, they just had to protect themselves from future suits. I had no problem with the nondisclosure agreement.”

Kildall said that when the meeting ended, he felt there was an understanding that Digital would supply IBM with its CP/M operating system. That night, Kildall continued, he and his wife happened to be on the same commercial flight from San Francisco to Miami as the IBM group, and they all talked some more. The Kildalls were headed for a long-awaited Caribbean vacation. Kildall said he was told to contact Sams when he returned from vacation, “but I couldn’t get through. He had apparently been moved from IBM. It was like he had moved off the planet.” The next thing he knew, Kildall said, Microsoft and IBM were doing business together on an operating system.

Sams said it’s possible a member of his team who visited Digital Research that day flew back to Boca Raton, rather than returning with him to Seattle. He said he just can’t remember. But Sams said he is absolutely, positively sure he never met with Gary Kildall. “Not unless he was there pretending to be someone else.”

Sams
was
later taken off Project Chess and given another assignment. But this was toward the end of October, over a month later.

Kildall said it was Gates, in an interview with the
London Times
after the IBM PC was unveiled to the world in August of

“It would certainly bother me if I had made that dumb a move, or if others had made it on my behalf,” said John Torode,

Kildall’s friend who designed the disk system for their microcomputer. He now runs a computer business in Seattle. “But how much preparation did Digital have? Did they know it was IBM that was coming? Did they have time to develop a strategy? Had they already concluded it was smart to tell IBM to go jump in the lake, or did it just happen because of the personalities of the folks who were there at the time? I’ve never discussed that with Gary. I never wanted to rub his nose in it.”

Regardless of what really happened that day, most of those in the computer industry believe Kildall’s actions helped make Microsoft the software giant it is today.

Sams said he subsequently telephoned Kildall after the ill- fated trek to Digital Research. “I told him we were serious, we really did want to talk with him. I had to assume we had gotten off to a bad start and that he wouldn’t arbitrarily refuse to do business with us, you know, did he have some religious opposition to us. He said ‘No, no, no, we really do want to talk with you.’ ” But Sams said he and others at IBM could not get Kildall to agree to spend the money to develop a 16-bit version of CP/M in the tight schedule IBM required. “We tried very hard to get a commitment from Gary,” Sams said. “When we couldn’t, I finally told him, ‘Look, we just can’t go with you. We’ve got to have a schedule and a commitment. We can get one from Gates.’ ”

In a series of meetings with Microsoft after the initial rebuff from Digital Research, Sams threw the operating systems problem in Gates’ lap. “This was the negotiating tactic we took with them,” said Sams. “We wanted this to be their problem, to find us the right operating system, one that we could integrate successfully on our schedule.”

Luck once again would shine on Bill Gates. An operating system for the 16-bit Intel chips had just been developed by Tim Paterson at Seattle Computer Products, not more than a twenty-minute drive from Microsoft.

Tim Paterson had always wanted to design an operating system for a microcomputer. But had the “Father of DOS” realized he was going to stir up so much controversy, he might have stuck with racing cars, which he enjoyed almost as much as programming.

After showing off Seattle Computer’s 8086 CPU boards at the National Computer Conference in the spring of 1979, Paterson had returned to Seattle to perfect the boards so they could be sold commercially. During this time, he was contacted by Digital Research, which wanted one of the CPU boards in order to develop a 16-bit version of CP/M. But Seattle Computer did not have any boards to spare. Paterson asked when Digital expected to have its new version of CP/M ready. By December, he was told.

Seattle Computer began shipping its first 8086 CPU boards to customers in November of 1979. Microsoft’s Stand-alone BASIC was offered as an option. These first customers were mostly software developers. By April of 1980, Digital Research had still not designed CP/M-86. Paterson decided to wait no longer; he would develop his own operating system.

“Here we had something that would work, but we were waiting and waiting for Digital to come out with their version of the operating system for the 8086,” said Rod Brock, owner of Seattle Computer Products. “They kept telling us any day now we will have it. This delay was really costing us sales. It’s hard to sell a product without an operating system. We were probably selling five to ten boards a month, but figured there were a lot more sales out there than that. We needed an operating system to get them.”

Five months later in September, around the time Jack Sams was being stonewalled by Dorothy McEwen and the lawyer at Digital Research, Paterson had his operating system up and running for the first time. He called it 86-QDOS, which stood for quick and dirty operating system.

Seattle Computer began shipping Paterson’s 86-QDOS to customers. “This was a real product,” said Paterson. “Everyone always thinks IBM was the first to have it. That’s crap. We shipped it a year before they did. It was used on our computer. We were selling a computer that was more than twice as fast as the one IBM was going to come out with.” (The reason for the difference in speed was that the IBM PC used the slower 8088 chip.) ,

Just as Gary Kildall has had to read over the years how he lost the IBM deal because he was off flying in the clouds, Paterson has had to read how he ripped off CP/M in developing an operating system that became the industry standard. Typical is this comment from an unidentified Digital employee who was quoted in a 1990
Business Month
article that depicted Gates as a silicon bully: “We never tried to patent CP/M. Nobody was patenting software then; it was almost unethical. But if we had, Microsoft probably couldn’t have developed MS DOS because parts of the original source code looked a lot like CP/M’s. How else did Paterson and Gates come up with that nice new operating system overnight?”

At one point, Kildall telephoned Paterson and accused him of “ripping off” CP/M.

“At the time,” said Paterson, “I told him I didn’t copy anything. I just took his printed documentation and did something that did the same thing. That’s not by any stretch violating any kind of intellectual property laws. Making the recipe in the book does not violate the copyright on the recipe. I’d be happy to debate this in front of anybody, any judge.”

Although Paterson’s operating system mimicked some CP/M functions, there were significant improvements. QDOS stored data on disk in a completely different way than CP/M did, and it also organized files differently. Paterson’s goal was to make it as easy as possible for software developers to be able to translate what had become a huge body of 8080 programs that ran on the popular CP/M so they could run on his operating system. He first obtained Intel’s manual for its 8086 chip, which had detailed rules for translating 8080 instructions into 8086 instructions. Paterson wrote a translator that followed Intel’s guide. He then got Digital’s CP/M manual, and for each 8080 function he wrote a corresponding 8086 function.

“Once you translated these programs, my operating system would take the CP/M function after translation and it would respond in the same way,” said Paterson. “To do this did not require ever having CP/M. It only required taking Digital’s manual and writing my operating system. And that’s exactly what I did. I never looked at Kildall’s code, just his manual.”

Once Paterson had 86-QDOS working, he contacted Paul Allen and asked him if Microsoft wanted to adapt any of its software for Seattle Computer’s new operating system. “That’s when they found out we had it,” said Paterson.

Up until then, Microsoft had been unsure what it was going to do about obtaining an operating system. Digital Research was out of the picture. IBM did not have time to develop an operating system within the 12-month deadline set by its corporate brass. Neither did Microsoft, at least not if it had to start from scratch. Without an operating system, the entire PC project appeared to be in jeopardy. “The feeling was if we couldn’t solve it, the project couldn’t go forward,” said Bob O’Rear, the Microsoft programmer who would soon be given technical responsibility for the operating system. “We’d have no languages to sell on the IBM PC. It was of paramount importance that we engineer a solution to the operating system equation. .. . We had to do something so that this project could go forward.”

In late September, Allen contacted Rod Brock and told him that Microsoft had a potential OEM customer who might be interested in Seattle Computer’s new operating system. Allen, who could not reveal the identity of the customer, wanted to know if Microsoft could act as the licensing agent. Brock said yes.

Gates would later say that obtaining Seattle Computer’s operating system saved Microsoft about one year of work.