Why did IBM choose DOS?
In 1980 a big computer company was needed an OS for the computer they had slapped together. I won't mention their name, but their initials were I.B.M. They didn't have time to develop the OS themselves -- so they hunted for a microcomputer OS they could stick in, that wouldn't be a threat to their important computers (their mainframes), and in came Bill Gates. This is that story. Why did IBM choose Microsoft DOS?
This is a common variant (though there are small differences) as told by few very common books: Accidental Empires, Hackers, or Pirates of Silicon Valley (TV Show). All the tales vary a little but the spirit is the same.
No offense -- but the commonly told story smells fishy on a couple of levels. It sounds more like what IBM or Microsoft said happened. As far as I can tell, that is who Cringley interviewed about the event. I have heard complaints (through the grapevine) from some over at Digital Research about that accounting of events and they've always claimed this was a distortion of the facts. Some of that is to be expected because they look stupid (blowing one of the biggest deals of the century) but a lot of it just doesn't make sense.
The first issue I have with this tale is Bill Gates. He is not a lamb that is afraid to go into other people's markets and he is not likely to turn business away (as history reflects). The story smells fishy to imply that he would turn IBM away at the door and only try to sell them Basic (missing an incredible opportunity). This is especially fishy to people who know his ruthless drive to win. So this telling is so completely out of character that one must really wonder. Was he really a fool (business-wise) and just lucky that IBM came back? I somehow doubt that.
The biggest problem with this story is that it is quite common to let your lawyers look over an NDA (and to give time for people to do so) -- especially back then when NDA's related to computers were so new. So IBM's complete lack of understanding of this issue can actually make them seem pretty stupid and fragile (more than I think they realize). But while this may not reflect favorably on IBM it may reflect better than the truth (or a less spun story).
What I heard and remembered
I was around during this time (and soon after) and was hearing some industry gossip. I did hear the fable (as later repeated in books) going around the tradeshows, but I had also heard some other intriguing stories and details as told in user groups and by insiders (or wannabe's) at that time.
I compiled the plausible stories into my own mental image of what actually happened. Of course, this is based on conjecture and gossip; but it is gossip that sounds far more plausible and in character than the marketing story coming from Microsoft and IBM of the time so I'm not sure I'm willing to suspend reason and disbelief enough to just buy their version of the story.
Here's what I heard, and remember:
An IBM Executive (Akers) and Mary Gates (Bill's Mom) were both doing things with the United Way. They got talking and it became known that IBM was looking for an OS for their new Microcomputer. Of course, Mary Gates took an opportunity to plug her son and all his accomplishments. She told Akers that her son Bill had an Operating System and that they should visit Microsoft and talk to him.
IBM has a policy of being thorough -- and they were going to check out all potential OS's themselves. But this wasn't their industry; they didn't know of all the potential OS's. They went to Microsoft and had their first meeting with Bill. Bill told them that he had been working on an OS for some time and could probably show them a demo in no time. As was a pattern for Microsoft throughout their history, they were lying. Microsoft quickly scrambled and started looking for something to show IBM for their next meeting. They borrowed Seattle Computing's QDOS. It was a rip-off of CP/M but IBM didn't know (nor did they care) who came first. And this version ran on an Intel Processor which IBM had already decided to use for other reasons.
IBM was interested -- but they had to dot their I's, cross their t's, and check everything else out. Unix was too resource intensive and they didn't want to make something that was so powerful that it would compete with their mini-computers, so it was put out of the running early on. IBM was getting into Microcomputers as much to use them as peripherals for their mainframes as anything else -- and they didn't want to overpower the Micros. IBM has a long history of playing games with product performance -- dumbing them down to "fit" the marketing scheme. Fear of challenging their mainframes was also why they chose one of the worst processors available the time (Intel) instead of one from TI, Zilog, Motorola, or other superior processors (or instead of making their own). They also chose the slowest variant (the 8088) instead of the superior 8086 for the same reason. Also, IBM had already used an Intel processor in another project they'd done: a smart terminal. They liked that it was non-standard and less popular so they could be different. And another bonus was that they were able to buy a piece of Intel (on the cheap) which they later sold (too soon) at a substantial profit.
Gary Kildall was supposed to have been somewhat a playboy or free-wheeling 60's left-over; a pot-smoking, free-sex kind of guy. This was the very late 70s (see 1980). Not only was his personality the opposite of IBM's but he is also rumored to have gotten caught having an affair (see free love) with an IBM executive's wife. This is a huge no-no at IBM -- and none too popular most places. This is why Digital Research was fairly quiet about the real story... but still complained about the accuracy as told by others. It doesn't look good for them to say "Huh-uh! The real reason we didn't get the contract is that our founder had the morals (and wisdom) of a Kennedy or a Clinton".
When IBM went looking for an OS, Digital Research (aka Gary Kildall) didn't have an ice-cube's chance in hell of getting the contract because of that indiscretion. He was blackballed, and had an enemy in IBM (the executive himself) and had violated their corporate culture ( a political death-sentence in a political company). This is compounded by the fact that another high-level executive was talking about "Mary Gates's boy Bill" (a golden child and a shoo-in). So the visit to Digital Research was strictly a "cover-your-butt" move and Gates had probably already promised them the world. I've never confirmed who the executive was whose wife allegedly had the affair with Gary but some of his family did chuckle at the retelling and didn't deny it.
While IBM did visit Digital Research and Digital Research probably did say "give us a day to have our lawyers look over the NDA". None of that was unusual. IBM used that as an excuse to bail. And then they later made the rather lame excuse about being slighted only because it was better than their even more lame original reason (that politics and personality were more important than technology).
The Boca-Raton boys (the ones in IBM slapping together the PC) were under extreme time pressures as well so they were being far less thorough than was normal and rushing. They also took a somewhat abnormal tack for IBM and figured they didn't really care about doing it right the first time. They were IBM; they could make mistakes and just fix them later(or that was the thinking).
IBM did allow Digital Research to sell their product for the PC. And CP/M 86 did make an attempt at the market. But it was too pricey, mainly because the development wasn't being sponsored by IBM and didn't get the IBM name (unlike Microsoft's DOS). So Gary Kildall and Digital Research lost out -- and Bill Gates lucked out.
Anyone that used both realized that CP/M was the far better and more mature product. Microsoft always had quality issues. But DOS was supported by IBM. And While CP/M was close, being that it was what DOS had ripped off in the first place, Microsoft had made it incompatible and it was still not DOS.
This is a compendium of the facts and tales as I've heard them, but again, I can't confirm much of it. And even then, it would be a bit of a he said / she said.
- at the time, Digital Research had sold 600,000 copies of CP/M
- Microsoft didn't have anything to show yet
- Yet IBM (a traditionally very conservative company) went with Microsoft, because of a guy missing one meeting?
That smells fishy. Something was going on. It was probably more substantial than Digital Research asking for time to review a Non-Disclosure Agreement and it was probably something political, and it left a lot of questions in my mind. The gossip about Kildall filled in the blanks just too well. It certainly sounds more likely (logical) than the "official" versions from IBM and Microsoft which just don't sound quite right.
Of course, the truth is sometimes more surprising than fiction -- and it is possible that both Gates (Microsoft) and IBM acted completely out of character. But I just tend to place my bets with Occam's Razor (or the stink test)... and the one told by IBM/Microsoft doesn't pass. The gossipy one sounds a lot more likely (to me).