Microsoft@50: The Windows Battle, with SVP for Microsoft's Desktop & Server Software, Paul Maritz

We're seeing another technological capability now that is going to introduce new entrants into this space. Whether it be another US company or the Chinese, I don't know. If I was at Microsoft, I would not rest. This is not a time to rest on your laurels. And there's nothing history teaches us. There's nothing that says that Microsoft of necessity has to be the force that it is in 15 years' time. In fact, if anything, history teaches us that eventually the center of tectonic plates shift and it's someone else who's on a different business model, different value proposition that gains traction.

Good to be with you today.

You know, it was on one level a very normal existence, on another level a very strange existence. White people lived cheek by jowl with black people but at the same time there was a great divide between the two. And over time, you know, that became more and more visible and obvious. And you know, I think a lot of people believed that eventually it was going to be unsustainable and a lot of people feared that it would probably end in a very violent situation, which in some cases like in the bush war in Rhodesia it did. But in South Africa fortunately thanks to remarkable leadership on both sides of the divide that was avoided. So even though people tend to kind of see South Africa as having a lot of problems today, compared to what it could have been, you know, it's obviously in a much better situation. So it certainly did impress upon me the realization that luck plays a big factor in life. The fact that I was born on one side of the divide and not the other side of the divide obviously had a huge impact on my subsequent life. And you know, I think over the years in the various roles I had, people who finally got sick of me saying to everyone, saying, you know, that we should be very humbled and thankful because for every one of us sitting in this room, there are probably hundreds or even thousands of people who are smarter than us, worked harder than us, but just for fortune, we're here and they're not. So I think all of us who were at Microsoft during the incredible journey from the mid-80s through to today have to realize that we were given the privilege to be part of something that occurs only a few times in history. So we should all be very cognizant of that.

Well, I had left South Africa partly because of this fear about what the future held and partly because I wanted to work in the computer industry. I was one of the first graduates of computer science in southern Africa and in the world. I graduated in 1977. You know, computer science was relatively new then but I knew I wanted to work in the computer industry as opposed to just using computers and that meant having to leave southern Africa. So I went first to the UK and started working for a company called Burroughs there. And if you can remember that name you're dating yourself.

And while I was there, I had originally thought I'd follow an academic career. So I'd gone back to university and had an offer to become a junior lecturer in Scotland. But my boss at Burroughs had moved to Intel and he persuaded me to come and join him in California at Intel. And I told my wife at the time, you know, we'll always have the opportunity to come back to the UK but probably this is our only opportunity to go to the US. So I went and you know, that again put me on a whole different path in life. So I spent five years working at Intel which was a very formative experience for me. Only afterwards I realized how good a management team Intel had in those days in Andy Grove and Bob Noyce and Gordon Moore, and they obviously thought very deeply about what leadership and management meant. And actually I brought many of those lessons to Microsoft later. Some of which became somewhat controversial things like ranking and rating. That all came from Intel to Microsoft.

Yeah. So it turns out that my kind of mentor, a guy by the name of Jim Harris who had brought me to Intel then went to work for Microsoft. He was the first head of OEM sales at Microsoft and I nearly joined in 1983 but I had startup fever. I always wanted to do a startup so I declined it, which I then subsequently rue because it meant I missed the initial IPO at Microsoft. And in '86 I tried to do a startup on my own and you know, as Jim Allchin put it, when you can no longer afford the milk to go with the cornflakes you have to be humble. So I put my tail between my legs and came back to Microsoft and took a job there in late 1986.

I had by then, as I said, I interviewed in 1983 and met Steve Ballmer and Bill Gates and BJ Bashi and others. So I had connections there. Jim Harris was still at Microsoft. A lot of people don't remember him because he left fairly soon after 1986, but he's actually the origin of the 'eat your own dog food' story.

We can come back to that. So I joined in 1986 kind of with tail between my legs and what they did is they said okay you can take over Xenix. People forget that Microsoft before MS-DOS even was the first commercial licensee of Unix from AT&T. In those days AT&T was still under the consent decree with the government and they were not allowed to go directly into the software business. So they had this asset called Unix that they wanted to try and market and they marketed it on a peculiarly strange licensing scheme. So I was given this 12-person group at Microsoft and told to do something with this business which I think was like a million and a half dollars a year. But by that stage Microsoft had already embarked on MS-DOS and signed the joint development agreement with IBM to do OS/2. So this was always going to be a backwater but that's what I was given. And so I eventually figured out a way of exploiting AT&T's licensing terms in such a way that they wrote their licenses as perpetual licenses. There was never any expiry. So when I figured out the loophole, there was no way out for them. So they eventually agreed to buy us out in effect and in return for a $15 per PC royalty anytime any form of Unix was sold on a PC. This is pre-Linux days. We would get $15 a year, that brought in like a million or a million and a half dollars a year to us. I always felt good that whatever else had happened I'd paid for my salary at Microsoft.

But then, you know, it was clearly a backwater and so I asked for something more ambitious. So I was given the networking business to try and compete with Novell. Remember in those days, it's 1987-88.

That's correct. So the product we formed was LAN Manager which used OS/2 on the server and the SMB networking stack etc. But you know, it was a tough road to hoe because Novell had 80% of the market. And that's actually the origin of 'eat your own dog food' because the phrase came from Jim Harris who always used to lean back after their presentation and sort of say in his big booming voice to everyone assembled, 'Yes, but will the dogs eat the dog food?' You know, as the final test. So we had, because we were nowhere in the networking business, I sent an email to the group saying look we've got no customers we're going to have to be our own beta test users. So Brian Valentine set up a server called slashdogfood. That was the first test server in the group and somehow things went on from there.

So you know, it took some time but finally, you know, Microsoft crushed Novell.

Well, it was Windows NT. I mean what killed Novell was finally the fact that Windows NT became a force on the server. And then the question was why have this other strange server that just serves files and print when you can have a general purpose server that does file and print and several other things and email and the rest of it. So we didn't take major market share. We kind of won by changing the ground rules, changing the nature of the game. You know, it's the same way that the applications group won against WordPerfect and Lotus 1-2-3 because they went to Windows and they changed the game. Until then, they could make no headway against WordPerfect and 1-2-3.

Yeah.

That happened in, I want to say 1990 and maybe, sorry, '89. It happened in 1989. Steve had wanted to move to... well, Steve was having to manage the IBM relationship which was becoming a real pain in the butt and time sink. So he handed it to me and I was given OS/2. And you know, trying to figure out what could be done. You know, in those days it was a sort of article of faith that the reason that Microsoft existed was because of the relationship with IBM. You know, in the late 80s, '87, '88, '89 we used to have these executive offsite retreats and there was always one of the topics was what do we do if IBM divorces us? You know, that was sort of the doomsday scenario. So the feeling was that we had to maintain the IBM relationship at all costs but at the same time OS/2 had become this beast that was so overburdened and trying to serve so many goals that it was just clearly failing as a system. And at the same time, you know, David Weiss and others had figured out how to extend Windows to handle larger address spaces initially on the 286 and then on the 386. So there was this increasing... Bill was getting very frustrated with OS/2 in particular and indirectly with the IBM relationship. But he didn't want to let go of it but at the same time he could taste and feel Windows that was much better suited to the 386 generation of PCs at that time. So there was this increasing tension. So I was given this job of trying to make something out of this and struggled at it valiantly for nine months. I remember there's a time in my life when I used to get on a redeye flight on Thursday night, fly to Boca Raton, get there 9:00 in the morning, be in meetings with IBM, turn around and fly back getting back to Seattle at midnight. You know, I became a lifetime American Airlines flyer because of that. But at the end of... and then we launched Windows 3.0 and I think that was 1990 was it? I could be wrong here, a year or two off, but it started to succeed. It put even greater pressure.

Yeah. Windows started taking off and it became, you know, the cognitive dissonance just got bigger and bigger as to why we were putting effort into these two directions. And the OS/2 team internally at Microsoft became more and more demoralized because they were having to work with IBM. They felt they had one hand tied behind their back while their colleagues at Windows were running. So eventually I came to the conclusion it was unsustainable that Microsoft and IBM had caught themselves into a situation where rather than getting the best out of both companies we were getting the worst out of both companies. So I eventually said to Bill and Steve that unless something changes I was just going to untie my hand. I was going to stop working with IBM unilaterally. I don't think they really believed me. But you know, I gave them a deadline in sort of two weeks or something and the day it came and went and nothing changed so I told the team to stop working with IBM. We just cut off all communication and stopped replying to emails, didn't do the joint build etc. Of course that caused the shit to hit the fan and so all of a sudden there were urgent meetings going right up to Bill Gates and Jim Cannavino who was the IBM senior vice president on the other side. And it precipitated a lot of brinksmanship. And here I have to hand it to Bill. People forget that IBM never paid a royalty on MS-DOS. The original deal in MS-DOS is they had a one-time fee, I think it was $50,000 or something, and after that they got MS-DOS for free. Or eventually IBM came back and said, 'Okay, we'll let you continue to market Windows if you give us Windows for free as well, so we can have both DOS and Windows for free.' And Bill refused. He said, 'No, I'm not going to do that.' And then they said, 'Okay.' And you know, so IBM said, 'Okay, in which case we're going to bury you. We're going to take OS/2 and we're going to compete directly with you and take you out of business.' So we had to decide obviously what to do. I had this investment in the OS/2 team and by then Dave Cutler and his team had joined the company. We decided what we would do is essentially walk away from OS/2 and redirect all of our so-called advanced operating system efforts towards a Windows API. And we came up with a strategy of having a common 32-bit API that would sit on top of Windows 4.0 and on top of what became Windows NT.