GAME OF X V.2

“If you’re going to look up evangelist in the dictionary, of course they’re going to give you an ecclesiastical definition, but if they gave you a tech definition, they should have a picture of Alex there.”

“We were kamikaze nerds, damn it. We’d throw ourselves on the grenade. We’d do whatever it took to win.”

The person who probably jumped the most grenades in Microsoft history had to be Alex St. John, a man who embodied genius, audacity, tenacity and foolishness almost in equal measure, along with a prodigious talent for disruption. With Craig Eisler and Eric Engstrom, he waged a war inside of Microsoft that changed the course of the company and, in some very real ways, made it possible for Xbox to happen.

St. John grew up in Alaska in an extremely rustic setting. He got his first exposure to computers and programming hanging around the University of Alaska’s library. While his father was on campus working on his PhD, St. John was “playing games on a 300 baud teletype and burning through reams of printer paper playing Super Star Trek, while the graduate students struggled to get their early chess programs to work by loading boxes of punch cards into the punch card reader.”

Eventually, St. John’s family moved into a log cabin that came complete with electricity, which he observes “greatly facilitated the use of a computer.” He received a Commodore VIC-20 for Christmas when he was 14 and immediately began copying game code from tech magazines. His first original game was what he calls a 1981 version of Diablo, with a single-pixel protagonist, “randomly generated dungeons of lines, asterisk treasures and scary single red pixel monsters.” Living without television through the long Alaskan winters left little to do but read and program.

When he was 16, St. John’s father moved to Amherst, MA, taking his son with him, but St. John, even as a teenager, was furiously independent and headstrong. “I got dropped into a public school, hated it, dropped out, and hitchhiked back to Alaska where I lived in the woods for a year.” Despite his lack of formal education, St. John was able to test his way into an electrical engineering/computer science program at the University of Alaska, but after getting married at 19, he dropped out of college to satisfy his new wife’s desire to be closer to her family in Boston.

The first thing on St. John’s agenda after arriving in Boston was to find a job, but he had issues. He describes his personal self-assessment and world view after entering what he calls “civilization” as, ‘I’m undereducated, therefore I must be stupid. I have to compensate for that somehow.”

At St. John’s first job interview with Linotype Hell Graphics, one of the interviewers asked him what programming languages he knew, to which he answered that he knew 14 programming languages, listing them out one by one. The incredulous interviewer then pulled out a 200-page book, slid it across the desk and asked, “Well then, how long do you think it would take you to learn PostScript?” After thumbing quickly through the book, he answered, “Well, it’s a dictionary based language like Forth… a week.” The interviewers were, to put it mildly, skeptical. “I remember they looked at each other and they laughed, like isn’t this kid funny, but I’m glad he thinks that. ‘Sure kid. Knock yourself out. Show us what you know in a week.’”

When St. John returned a week later and announced, “I know it,” the manager he met with regarded his statement with the expected skepticism, but agreed to give him a job working under their resident PostScript expert who had been trained at Adobe and had been using PostScript for years. The idea was that St. John could not possibly know PostScript, but that he might be capable of learning from their expert.

Working under the “expert” didn’t go according to script—PostScript or any other script. It turns out that the “expert” had some deficiencies. “He was constantly struggling with PostScript, and he’d call me over to show me a problem, and I’d go, ‘Oh, you’ve just got to do that,’ and fix it.” After a couple of weeks, St. John began to struggle with that familiar world view of inferiority that he carried around. How was it that he already knew how to work with PostScript better than a guy who’s really “seasoned and experienced”?

As it became clear that St. John knew what he was doing, he was placed in a customer support role for a PostScript RIP (raster image processor) clone that had been licensed from a company called Hyphen and then developed around an Apollo UNIX workstation by a team of engineers in Kiel, Germany. The problem with supporting customers of this million dollar product was that, in St. John’s words, “It was a piece of shit… just worked like crap.”

St. John tried to work with the German engineers, even pointing out the major bugs in their code, but they refused to recognize any shortcomings in the software. In a German accent, he quotes them: “The product is not designed to operate outside those specifications. Tell the customers they are not using it as required. It is not supposed to support that.” He might as well go back to the customers and tell them, “The Germans say fuck you, but thanks for the millions of bucks…”

Of course, this translated into job insecurity. He was thinking “Oh my god. It’s broken. I can’t make it work. If I tell the customers I can’t fix it… I can’t do anything about it… I’ll be fired. They’ll find out that I must not be smart enough to solve these kinds of problems because I don’t know how to deal with this.”

St. John’s solution was to take the broken jobs his customers sent in and spend his nights recoding them. “I would modify the code in the job to work around the bugs in the RIP manually and then hand the job back to the customer, rendered, just to make it work.”

One of several problems to this approach was that it meant little to no sleep. To keep his job, St. John was working day and night. The other problem was that he was the only one getting results. The other support people were getting destroyed. And so his manager came and said to him, “Hey Alex, we’re sending you to New York. They’re impressed as hell. Everybody demands you at this point. They want you to drive to New York to train all the other support people to do what you do.” At which point St. John was forced to admit that there was no way he could train people to do what he did. “Rick, I’m fixing the fucking postscript jobs,” he said. “These poor people can’t code at all. There’s nothing I can do.”

At this point, desperate for a solution, he told his manager, “You know, there’s another group in the company that’s using the PostScript RIP from a British company called Harlequin.” Then he added, ‘It’s a piece of shit, too, but maybe they can fix the bugs.”

“So I call up this little English company that also had a shit RIP and got in contact with Jeremy Kenyon, who was running that product, and I said, ‘Look. I’ve got this many of these RIPs out here. How much do you sell them for?’ And he says, ‘Fifteen hundred bucks each.’ Nothing. I could give them away. So I said, ‘Listen, if you can run these jobs that I’m going to send you and send me RIPs that run them, I’ll buy them from you at fifteen hundred bucks a pop.’ And he agreed. And so I’d get a broken job, I’d send it to Harlequin, they would fix a version of their RIP according to my specifications, that ran the job the way it’s supposed to, and send it to me. I would replace the German RIP on the Apollo workstations with a Harlequin RIP, and the jobs would magically work.”

“It was a secret. Nobody but my manager and I knew I was doing this. By the time I got to Microsoft, I was doing this on purpose. But this was the first time for me, and I was completely ignorant, which is why it’s so funny. I was terrified every day that I was going to get found out and fired.” Although this was the first time, it was definitely not last time that St. John would come up with a brilliant, but definitely not kosher, solution to a problem.

The relationship with Harlequin was working so well that soon they had St. John writing the product specifications for them. One of the ideas he came up with was something he called a “parallel throughput pipeline,” which he modeled in a spreadsheet. The engineers at Harlequin followed St. John’s instructions and produced the parallel processing software he had designed. St. John then began inserting it into the broken systems. And something happened… “Not only was I replacing the Hyphen RIPs secretly and everything was working magically now, but the performance went through the roof—something like ten times the productivity with this new architecture I was slipping in there. And suddenly these things were selling like crazy.” St. John’s little program, which Linotype considered a low-end product, was out-performing the million dollar systems and beginning to drive them out of business.

St. John had no idea what he was doing, nor did he yet realize what talent for disruption he possessed, but his little “low-end” parallel RIP processor completely disrupted the PostScript printing world. “I hadn’t been there that long… maybe six, seven months, and my manager was handling all this and kind of shielding me from it, so I only imagined the conversations that took place. ‘Listen, you guys. I gotta tell ya. I got a freak of nature kid here who is just building the shit and replacing it and he designed this product, and we’ve been replacing your German shit with something over here, and that’s why it all works. And the kid doesn’t know what he is.’”

And within a few months, everybody in his entire office was laid off, except for him and his manager. “I remember rollerblading around the empty cubes in my office, just stupid-go-lucky, while my manager must have been sweating bullets in his chair, going ‘Oh my god. I’d be out of here like everybody else—this late in my career—if this fucking kid hadn’t saved my ass.’”

The president of Linotype, with several executives, flew in to meet with Alex and offer him a job in Germany. “I’m trying to remember the phrase, but it was basically like, to work for the Germans so that they could teach me the ways of building great products. The same Germans I hated. Just hated. And so the proposition was, ‘We’re going to pay all your expenses to move to Germany and work for these guys to help develop…’ It was portrayed to me like a ‘you’re going to learn from them’ kind of thing.’ At the same time, I’d embarrassed the crap out of them. I made them look so bad, so whatever politics was going on, I definitely made a mess. So they were trying to persuade me to move, and I resigned. I said, ‘No. I’m outta here. Goodbye.’”

Shortly after quitting Linotype, St. John landed a job with SciTech, who wanted to send him to Israel, but Harlequin’s founder and CEO contacted him and made him an offer that he couldn’t refuse. He would move to Cambridge, England and run their PostScript RIP development. And he didn’t know it yet, but while he was at it, he would help destroy Adobe’s high-end RIP business. “El cheapo off-the-shelf Macs and PCs were blowing away their sixty and hundred thousand dollar proprietary hardware RIPs. So Harlequin did enormous damage to Adobe’s own high-end business before Microsoft recruited me. The funny thing, looking back, at the time it was all a great adventure and I didn’t know what was going on, but Microsoft… They definitely realized what I was, and then it was probably about that time that I realized, oh wait a minute, I’ve got some secret sauce here.”

St. John recalls that his time at Harlequin was both enlightening and amazingly fun. His desk was in the entry hallway of Barrington Hall, a great manor house set in the countryside where “you’d be sitting there coding and there’d be deer and bunnies running around in the fields.” Among the leisure time activities were playing croquet, juggling flaming torches, laser tag, and underground larping—having mock battles “beating the crap out of each other and running around in the dark” with latex weapons in the Chiselhurst Caves, which ran for miles and were “dangerous as all shit.” He quips that “it’s completely illegal to have that much fun in the United States.”

But with all the fun, St. John gained enormous respect for the British engineers and their math skills, while gaining a good deal of notoriety as a contributor to periodicals like Color Publishing and Seybold Magazine. “The products had become famous in the RIP community, so I would write for these tech magazines, like I do now, on printing and publishing technology and color science and screening because I was an expert on all that stuff.”

St. John left Harlequin in 1992 and went to work for ECRM, a high end laser image setter company in Boston. He had applied for a job at Adobe, but was told that he was overqualified for any jobs they had in Boston. His consulting work was paying well, but he was getting bored. Then, in late 1992, he somewhat reluctantly gave in to a recruiter from Microsoft and agreed to an interview in Redmond, against his wife’s wishes; she wanted to stay on the East Coast.

He knew that Windows sucked as a publishing platform. On the other hand, he admired Microsoft, which he viewed as “the scrappy upstart challenging the evil corporate monopoly, IBM…” So he agreed to come for “a free all-expense-paid vacation to Redmond.”

“I arrived in Redmond in December dressed in shorts and a t-shirt for an interview I had little interest in and knew nothing about. They put me through the famous Microsoft interview grinder of that era: eleven interviews in rapid succession each with a blend of different technical, logical and personality challenges.”

Ken Fowles, who would become one of St. John’s best friends at the company, asked what the ideal remote control would be. St. John suggested something with one button and maybe a thumb rocker switch. Cameron Myhrvold asked some questions about PostScript, and Brian Moran (while using a Bowie knife to pick his nails, although later Moran claims that it was probably a Spyderco), asked some technical questions about debugging protected mode and DOS interrupts.

All the while, St. John had no idea what job he was being interviewed for, and his confusion was actually reasonable, because the people who were interviewing him didn’t know, either. “First and foremost, you were taught to look for talent and ability, and then figure out where we fit the guy into the company,” said Cameron Myhrvold. And with St. John, he said, “There are people who understand things, and there are people who are able to explain them. And so one of the things in looking for evangelists is you needed guys who were articulate. You needed guys who were compelling. You needed guys who were good storytellers. And he absolutely could do that. All I had to do was start peppering him with questions about Harlequin and their business and Adobe and PostScript, and it became very quickly obvious that this guy had a remarkable handle on the technology, but probably even more impressive was how articulate he was in explaining it.”

St. John left the interviews with a job offer and a proposed title—Senior Manager—but no idea what he would be managing—and stock options, which he discounted in the belief that Microsoft’s rapid growth had slowed. Because his wife didn’t want to move, he was still thinking of saying no.

That evening, he got a call from Ken Fowles, who invited him to come to Microsoft’s Christmas party, which took place at the Seattle Convention Center, and to bring the family. “There were huge lines of people checking in. First people checked in at security, then they checked their coats, then they checked their children and were issued beepers coded to bands put on their kids who were subsequently ushered away by clowns to a huge circus they had set up in one of the halls.”

Microsoft had taken over the entire convention center, and the party was on a scale unlike anything St. John had ever witnessed. He was told that there were probably 15,000 people in attendance, half of whom were millionaires from their Microsoft stock. “I confess to being pretty awestruck,” he wrote later. “I didn’t care what they wanted me to do, I just wanted to be part of it… and Ken Fowles had just given me my first lesson on what a Microsoft evangelist was supposed to do.”

So St. John convinced his wife to try it for a couple of years and joined Microsoft to become, he eventually discovered, a publishing technology evangelist for the Developer Relations Group.

St. John’s talent for disruption did not abandon him at Microsoft. In fact, it flourished, and it only took three months on the job to make an appearance. It all began when his boss, Brad Struss, whom St. John described as a tall, soft-spoken kid with seemingly infinite patience, handed him the phone one day. On the other end of the line was a reporter from the very influential periodical Info World. Everybody read InfoWorld in those days, including Bill Gates, but nobody had briefed St. John on talking to reporters. Struss told him that a reporter wanted to talk to a Microsoft executive about Windows as a publishing platform. “I didn’t know at the time that one of my jobs was playing ‘Microsoft Executive’ as needed. Many evangelists had Microsoft business cards printed with a variety of titles they could deploy as needed in their various endeavors. I had no experience dealing with hostile media before so when I took the call I simply answered the reporter’s questions honestly. The next day when I showed up to work I was greeted by Brad wearing his trademark ashen ‘What have you done to me?’ expression.”

St. John opened his email that morning to a guaranteed career death sentence at Microsoft—an angry missive from Bill Gates himself, ccing every senior Microsoft executive, including Brad Silverberg, Paul Maritz, Cameron Myhrvold, his own managers and the head of Microsoft PR. As St. John remembers it, Gates wrote, “Where do you get off telling the press that Windows is an inferior publishing platform? You don’t know what you’re talking about and you shouldn’t be communicating with anybody in the media about it. Windows is a superior platform to Apple at publishing and you were hired to get that message out.” Gates had also attached the front page InfoWorld article entitled, “Microsoft Executive Acknowledges Windows is Inferior Publishing Platform.”

Despondent, St. John decided to take a walk outside in the beautiful, parklike setting that was the Microsoft campus. He wasn’t seeing the beauty. All he was seeing was disaster. What was he going to tell his family? He had just signed a lease. What would he do about that? Would they have to return to the East Coast?

After a time, his thoughts and emotions began to change. “I think it took me a good hour of walking to work past denial, and deep into despair such that by the time it was starting to get dark and I felt compelled to return to my office if for no other reason than to pack my desk. Rage had begun to set in. F**K IT! If I’m fired anyway I might as well respond to Bill G’s email!”

Bill I’m really sorry if I let you down, I really screwed up, I had no idea how to handle that situation and I said some really stupid things. I absolutely understand if this is my last day at the company. Obviously even if I felt that our publishing technology was obviously inferior to Apples I shouldn’t have acknowledged it outside the company.

That said, you don’t really believe that Windows is competitive with Apple do you? I don’t know who you’ve been listening to but I absolutely know what I’m talking about and Windows is a complete disaster as a publishing platform compared to Apple and absolutely everybody in the industry you hired me to represent knows it. You don’t have a single “publishing” expert here at the company; every single person I have met related to publishing in the technology teams is little more than a 300dpi HP laser printer expert. My credibility with that community wouldn’t stand a snowballs chance in hell with that community if you expect me to take the position that Windows is great when everybody but apparently you KNOW otherwise.

St. John then proceeded to list all the specific problems Windows publishing had in comparison with Apple’s.

This email prompted a response, but not necessarily the response that St. John had anticipated. He still fully expected to get fired, but in a way he didn’t care anymore. He was relieved. Why pull his punches if he was going to be gone anyway? His relief was short-lived, however. Gates had fired off another email, this time saying, “This new guy we hired to be our Publishing Evangelist tells me you people have been lying to me all these years about how great our publishing technology, which has been responsible for our abysmal performance against Apple in this space all along. I’d like to know how the issues he’s raised are being addressed by your teams going forward.”

The reaction was swift and furious. “Imagine if you will the collective ire of dozens of senior Microsoft Executives with thousands of years of industry experience and credibility and huge teams of the most elite, highest IQ engineers Microsoft’s rigorous recruiting standards could hire getting called ‘stupid’ and/or ‘liars’ in front of Bill Gates and having Bill apparently agree. I got POUNDED.”

Brad Silverberg, who would soon become one of St. John’s senior benefactors, had a decidedly strong reaction to the situation. According to Cameron Myhrvold, “Brad called me up and he’s like, ‘Cam, who the fuck is this guy?’ And I was like, ‘Here’s his background. He hasn’t been here this long. I think he really knows his stuff.’ So he says, ‘Well I’ve got to meet him. Get him over in my office.’” When they met, according to St. John a wild-eyed Silverberg said, “’If you ever speak to the press or Bill Gates again about my product without first discussing it with me, so help me…’ or something to that effect.”

Not for the last time, however, St. John’s management team and DRG colleagues rallied to his defense. He continued regular email correspondence with Gates and over a period of several months of shouting matches and recriminations, he proved, again and again, that he knew what he was talking about. As he put it, “I had somehow clawed my way from the bottom of the pile to the top.”

Along the way from the bottom to the top, St. John went head-to-head with some of Microsoft’s top printing experts, including Dennis Adler and David Snipp.

St. John expresses nothing but admiration for Adler. “I think of Adler as being in some sense the Microsoft employee archetype. Incredibly smart, incredibly confident, incredibly determined, and absolutely no bullshit. In terms of somebody that got shit done, he was that way.” St. John’s admiration wasn’t reciprocated at first when he completely disrupted the publishing group’s schedules and caused him personal embarrassment. So Adler, along with Eric Bidstrup, pushed back, but, as St. John explains, that was what he was expected to do as his domain was being challenged. “Microsoft arrogance aside,” observes St. John, “in that environment I think you had to be that arrogant to survive. So I forgave people for the arrogance because they got beaten on from every direction. So if you weren’t tough, you couldn’t make it. He was a tough guy.”

Adler was tough, but practical. This was another trait that St. John recognized and admired. “If you had the winning argument, they would adapt. They’d usually do it very efficiently. They do the most minimal thing they had to. If you won an argument rationally and logically, then they would do the most logical, rational, minimal thing they could do to accommodate your correct point of view. You could say that I came in, surprised them, threw over their apple cart, would not be stopped. Would not be shut up. No amount of pressure on me to shut the hell up, no amount of ‘we’re too busy. We don’t have time. We don’t have resources.’ I got past all of that with them, and ultimately, they went and did the Adobe deal. They involved me in all of the planning and reviews, and they overhauled all of that print stuff with a very tight schedule.”

Besides Adler, there was David Snipp—THE printing guru at Microsoft at the time—the one that everybody deferred to. And he was also in St. John’s way.

“I had to destroy his credibility to get anything done. And he was a nice guy.” Snipp was an undeniable HP 300dpi laser printer expert, but that was about it, and he was no match for St. John. Like many of the senior people at Microsoft, Snipp did not appreciate being challenged by this brash young man, and so he called a meeting intended to solidify his leadership in the printing division once and for all. St. John recalls that it was a huge meeting with people from both the Windows NT and Windows teams as well as all the significant people from the printing side.

“I decided I was going to go into that meeting and break him. Just break his credibility in front of everybody. What happened was that he was talking about PostScript, and he was going on… this PostScript and that PostScript, and I said, ‘You’re talking about PostScript level 1.’

“And I go, ‘You know the standard for PostScript is Level 2, and it changed in 1987.’

“I go, ‘All modern PostScript printers that are being manufactured are PostScript Level 2. The PostScript drivers that you are supporting and developing for Windows 95 and NT are based on 1987 PostScript standards. Have you even read the PostScript Level 2 book?’

“And I pulled out this massive red book, and I threw it across the table at him, in front of him and said, ‘This is the fucking PostScript Level 2. This is the standard from 1987, from Adobe, for PostScript. You had this entire company building obsolete printing architecture. Apple’s fucking PostScript drivers are based on this, and you don’t even know it exists?’

“And I just laid into him in front of everybody, and people’s eyes went wide. And I just ranted at him in frustration, going, ‘Dude. You go pick up the fucking manual. Here’s what it is. This is what it’s about. The drivers are 32 bit… duh duh duh.’ I just raved at him. And after that he was done. People quickly concluded, ‘That guy doesn’t even fucking know what he’s talking about compared to St. John. Whatever Alex says is gospel, and you fucking don’t argue with Alex, or he’ll destroy you in front of everybody.’”

Soon after that confrontation, Snipp left Microsoft, and although St. John says he feels bad about what he did to the man—”he was a nice guy”—he still believes he had to do what he did. “The nice thing is that Bidstrup and Adler, they just got on board and adapted. So, unlike Snipp, who couldn’t learn, they didn’t need to be killed because they got with it. Snipp was set in his ways, and that wasn’t going to change, and unless his reputation for being the expert got broken, nothing was going to change.”

St. John intimidated, bullied and out argued everyone until it became clear that he had won the Windows publishing battle. As he later recalls it, “I wreaked havoc inside Microsoft ‘influencing’ the product groups to redesign the Windows 95 and Windows NT graphics and print architectures. I brought Adobe in to strike a deal with Brad Silverberg to write the PostScript driver for Windows 95 in exchange for Adobe application support for Windows 95. I recruited senior product managers from Aldus to run the Windows NT print group. We cloned Adobe’s Type 1 font technology to kill Adobe’s ATM product (ruthless), licensed Kodak’s color calibration technology, persuaded every major desktop publishing application developer to port to Windows 95 and booked Bill Gates to keynote the annual Seybold conference. By 1994, the impact of the work had been so sweeping that I was quite possibly the first Microsoft evangelist to succeed himself out of a job. Windows 95 would have every major desktop publishing and media authoring product available for it by 1995. Apple would lose its dominance in media authoring.”

On the other hand, he adds, “In retrospect I never recall a day arriving when it occurred to me that I was not going to get fired the next morning and moderated my ‘F**K-You I’m out of here!’ attitude appropriately. It seemed to be working for me… Brad on the other hand seemed to age rapidly during the time he was my manager for reasons I could only imagine…”

Although St. John began a goat and emerged a hero, this was just the prelude to his career at Microsoft, and the best—and the worst—was yet to come.

“The key to winning a Microsoft technology debate was a mixture of being right, being confident and articulate, and being aggressive enough to force the issue but not angry enough to cause people to shut-down. You had to get that balance right.”

Alex St. John was learning on the job. In his put-down of David Snipp, he admits that he yelled because “I was trying to burn him out of his home… so to speak.” So even though it wasn’t truly personal, it must have seemed that way to Snipp. Over time, St. John refined his approach. “When you get ‘mad’ in front of people it suggests a lack of confidence and loss of control. When you yell ‘With a smile’ it’s scarier because you’re in control, you’re confident and you’re obviously enjoying yourself. This style of debate has the property of being charismatic and appealing to people. It allows them to change their positions because they WANT to be on your side of the argument, and you’ve made it clear that you aren’t really mad at them while you are arguing with them. It leaves the door open to resolution of the disagreement.”

According to St. John, this is an important “nuance” that printed accounts often miss, pointing out that it was a fairly common type of argumentative style used by the Microsoft’s executive leadership, including Gates, Ballmer and Myhrvold. “You hear all these stories about how they yelled at people and in print it looks very bad… the missing nuance is that very often they did this with a smile that suggested that this was just a debate among friends and that by challenging you they were inviting you to change their minds. That missing nuance is important to really understanding why these people were actually effective leaders and NOT simply bullies.”

“Of course you know I love being described as ‘vanquishing’ anything… Who doesn’t? But I’d like to think that a big part of my influence came from winning respect… sometimes by employing a certain measure of fear along with expertise and certainly by making myself impossible to ignore. You didn’t get the opportunity to earn important people’s respect at Microsoft if you didn’t cut through the noise in their day to become the issue they had to deal with next.”

On his relationship with Bill Gates, St. John observes, “While everybody was beating on Gates’ door for influence, I had no idea I had a lot of influence with him at that age. I just took it for granted that Bill listened carefully to what I told him. I was loud and brazen and wouldn’t shut up, and he seemed to pay a lot of attention to that in retrospect. But at that age, I didn’t think anything of it. I yelled at everybody.”

Of course, Gates was also known for his outbursts, and almost everyone I interviewed had witnessed at least one of them. And everyone has nothing but praise for the man. Business development specialist Chris Phillips recalls his experiences with Gates: “I’ve been screamed at by Bill, and been sworn at and told that I was an idiot, and blah, blah, blah, and all those things are true. He’d get pissed and be that way… when you had Bill engaged, even when he was yelling at you, you were like, ‘Wow. That’s insightful.’ If you could put aside your emotions and actually the bullshit that he was screaming about, you could get to where, ‘I hear his point.’ Especially I learned that in legal stuff, because he was a great legal mind. But all those things are true about him. A complex man. But when it came to strategy and seeing markets and stuff, Steve and he are pretty uncanny in that regard. Very few people I’ve met over the years can operate like that.”

James Plamondon: ‘Why don’t you go fuckin’ do games? You don’t have to be Mr. Publishing. The worst they can do is fire you, and you’re already up for that, so why don’t you just go do games?’

After his success on the publishing side, his manager offered him a new job, as the MAPI evangelist. MAPI was the Messaging Application Programming Interface, which was part of Microsoft’s mail messaging. “I couldn’t have been less excited about it.” The idea was that St. John would work on MAPI until after the Windows 95 launch, at which point the powers that be would figure out what to do with him next.

Meanwhile, Rick Segal was looking for a games evangelist. Once Windows 95 launched, there would be a huge need for evangelists because, basically, Windows 95 was going to break a lot of existing DOS games. Previous versions of Windows had actually been built directly over DOS, but Windows 95 was not. There was something called the DOS emulation box, but, according to St. John, DOS emulation would break probably 95 percent of currently working DOS games, which would, of course completely destroy the DOS games market, and with it, basically the whole concept of games for PCs if Windows couldn’t take up the slack. St. John describes the position essentially as going to the game companies and saying, “Go test your games on our new DOS extender, or you’re just going to get obliterated.”

St. John applied for the job, as did Craig Eisler. Eisler was already the tools evangelist, and certainly qualified. Eisler was working on, in St. John’s words, “killing Borland,” and he was the author/creator of Watcom’s 32-bit DOS extender, which was a tool that all games used at the time in order to run in 32-bit mode in DOS. “So he in some sense was already a game industry celebrity because he made the technology that everybody used to extend DOS, and was well known for that.”

On the other hand, St. John had proved himself unstoppable and persuasive in his Microsoft debut, and he had some knowledge of 3D graphics, which everybody knew was going to be very important. Eisler was already doing a great job as the tools evangelist, and St. John was clearly wasted working on MAPI. Also, “I think Rick also favored me because I was crazier than Craig, and he liked that…”

According to St. John, what flipped it in his favor in the end began with the announcement that Intel was developing its own multimedia architecture, including a whole family of sound and 3D multimedia APIs called Native Signal Processing (NSP). They asked Microsoft to send a speaker to their event, which caught VP Paul Maritz by surprise, again according to St. John, who states that Maritz needed somebody to go to the event and represent Microsoft. The mission, however, was really to find out what Intel was up to. Rick Segal tapped St. John for the job, telling him, “I’ve got to send somebody on a mission to stand on stage at this Intel event in two days and say some stuff that sounds supportive without looking foolish, and I need to find out what Intel is doing.”

So St. John’s real job was to study Intel’s product and write a strategic analysis for Maritz, which he did, saying in effect that Intel was attempting to usurp control of Windows media architecture with software emulation. In his assessment, he said that Intel was “trying to completely kill the market for hardware and hardware drivers that offload processing from an Intel CPU. It will all be on the CPU. Everything was software emulated and cumbersome. And I recommended that the counter strategy to that would be for Microsoft to provide much more robust driver support for proprietary hardware acceleration, thereby fostering a market for competitive proprietary hardware that outperforms Intel’s solution…” In essence, “If you want to prevent Intel from being able to replace all of your shitty multimedia features with software emulation, which is possible because they are so crappy and primitive, the way to do it is make them sophisticated, create lots of hardware vendors that depend on it, and then Intel will have a hard time replacing it with shitty software emulation.”

(Under pressure from Microsoft, Intel ultimately abandoned its NSP initiative, which in turn ultimately became part of the famous antitrust suit against Microsoft, with assertions that Microsoft used its considerable influence unfairly to force Intel to back down. Microsoft’s lawyers, meanwhile, asserted that Intel abandoned NSP because it was an inferior solution. (http://www.businessweek.com/microsoft/updates/up81110a.htm)

St. John credits his actions around NSP with getting the job. He also describes his rival as “a machine and a human calculator and a genius,” then he laughs and states, “but his charismatic skills are a little more limited.” But Eisler was not just a rival, he was a friend. Along with Eric Engstrom, the three regularly got together in the mornings to lift weights and shoot the breeze. In the aftermath, Eisler’s relationship with St. John became somewhat chilly, but only for a while.

The job St. John had competed to win wasn’t an easy one. He had to go tell the game companies that they needed to test their games against the “DOS box” to see if they would be compatible with Windows 95. Of course, he really wanted to get them to make Windows games, but he knew how utterly crappy Windows was for games, so that wasn’t really a viable option. And besides, he was from Microsoft. “I’m here to help you,” just didn’t seem to have the desired impact. In Alex’ words, what they heard was, “And yeah, we’re just going to destroy your little DOS world that you were able to live in while evil corporate blue giants took over the enterprise space. You were able to cozen down here in DOS and ignore us all. I’m here to tell you that I’m just destroying your world, and I want your support for that.”

At a meeting at Mindscape, people began doing Darth Vader impressions and heavy breathing while he spoke. This reception became common enough that he adapted to it. “I’d come in with a smile and my business card and I’d say, ‘Hello, I’m from Microsoft. I’m here to help you,’ and get a laugh out of that, which became a necessary ice breaker. But I mean, I was the herald of doom for these guys. (no pun intended).”

Meanwhile, at Microsoft, there was an effort underway to fix incompatible DOS games. It was basically the work of one man. “One of the fascinating things about Microsoft at that time is there were people there who were giants. I’ve never met their like since. Their intellectual capabilities surpassed anything I’ve met in recent times. I know they’re out there still, but Microsoft had an amazing concentration of them. Chen was a little like an Asian Craig Eisler in that he had this sort of infinite capacity and patience for minutia.”

The man St. John describes is Raymond Chen, who was given an insane mission. “So what Raymond Chen would have to do was, I’d bring these games in… I had boxes and boxes of them. Sometimes all the game developer could do was hand me a box and say good luck. We can’t do anything about it. It’s done. And Raymond Chen would have to write patches for each of these games in assembly language in order to get them to run in the DOS Box. There were thousands of them, and they were enormously complex, and it was very low level stuff, and he did this. One of the fascinating things—one of the things that won a lot favor in the game industry for Microsoft was the contact between myself and Raymond Chen and these early developers when they saw what this guy did to make their games work. Because he literally did heroics. Just mind staggering. You just go, that’s… not only is it insane to do, nobody rational would ever undertake something incredibly tedious.”

Another fan of Chen’s was Origin Systems’ director of technology, Zach Simpson. During that same period of time, Simpson had come to Microsoft to spend a month, along with Tony Braden and Frank Savage, working on Windows interfaces for just about everything, from joysticks to sound and graphics. He talked about how Chen fixed Ultimas 7 and 8.

“Ultima 7 was a DOS game. So was Ultima 8. Ultima used a horrific hack called Big Real Mode where you kicked the processor into this kind of strange mode that nobody used, and it was incompatible with the EMS managers, so it was a real nightmare of people calling in, trying to disable their EMS managers to get it to work. In Ultima 8 we used a 286 DOS extender that we hacked to be a 32-bit DOS extender, and that was a little more compatible with EMS and other DOS management at the time. Ultima 7 and 8 were before Chicago (Windows 95), and Microsoft paid a guy to go through the entire catalog of popular games in the world and make sure each one would run in DOS compatibility mode under Chicago, which is like the most horrific job I think anyone’s ever foisted on somebody. And so that poor bastard got Ultima 7 to work, and we brought him flowers to tell him we appreciated his work. He was like a hero, and so you could still play Ultima 7 and 8 under Windows 95.”

“Microsoft was just paranoid about QuickTime, and multimedia in general, because it was clearly the way computers were going, and Microsoft just wasn’t able to dislodge QuickTime as the de facto standard.”

Bill Gates always had something to worry about, and Apple was always high on his list. QuickTime was a good example of what concerned Gates. It was a multimedia technology that played videos on the Macintosh, and it did so better than any other video player at the time. In part because of Quicktime, but also because the Mac was a native windowed GUI operating system built into the hardware, Apple was dominating the emerging multimedia space. Microsoft’s Windows was not as friendly to multimedia. Windows was built entirely in software that ran on top of the same hardware architecture that had originally run DOS. In the world of multimedia in the mid-90s, Microsoft was playing catch-up—and never quite getting there.

Of course, Microsoft was trying to make the best of the situation, including an attempt to play “let’s make a deal.” James Plamondon recalls a letter sent by Microsoft’s president, Mike Maples to Apple with an offer. Maples’ offer was that Microsoft would cede video technology to Apple, agreeing to make QuickTime the video standard in Windows (which would have cost Microsoft beaucoup bucks) in exchange for Apple abandoning its OpenDoc in favor of Microsoft’s Object Linking and Embedding (OLE) technology. According to Plamondon, it was a case of “We’ll give up our innovation if you give up your innovation,” and it might not even have been legal. Apple did not take the deal.

There was even a video solution called Tiger that Craig Mundie was leading. According to Jason Robar, Tiger was a plan to serve up videos to people in their homes. One part of the plan involved creating a video server farm. Another aspect of the plan involved having NT Server in everybody’s household, residing under their TV. But, as Robar puts it, “You’re going to buy a nice $1500-$2000 computer to put underneath your TV, and everyone’s going to do that because doesn’t everyone have an extra $1500 to spend on TV?” Robar further observes that the people behind this idea were “guys who had already had huge stock options pay out.” In fact, the head of the project was an audiophile whose speakers cost $20,000 each. “So his viewpoint of the consumer is very skewed.” It was a visionary plan, but way ahead of its time, and of course it was never implemented, despite being in planning and development for several years.

To people like St. John and Plamondon, it was clear that attacking QuickTime directly was like, in Art of War terms, attacking a “fortified city”. To attack the fortified city was sheer folly. “That’s where the enemy had successfully predicted your attack and concentrated his forces to defend against it,” said Plamondon. “However, strengthening the city’s defenses necessarily meant weakening them elsewhere. The trick, then, to is get the enemy to fortify their defenses in Location A, and then attack in Location B.”

So the question was, where do you attack if you want to win the battle? Not many people at Microsoft would have come up with a very viable solution, let alone one that seemed, at first glance, to be utterly loony, but when Rick Segal asked Alex St. John what he thought they should do about Quicktime, his answer was “Go after games.”

According to St. John, Segal was in charge of “fucking up Apple’s multimedia strategy”, and he had a team of media evangelists whose primary purpose was to destroy Apple’s perceived multimedia leadership. So when St. John suggested that they focus on games, where Apple was not focused or successful, Segal, who had no problem supporting crazy initiatives, said, ‘Go for it.” St. John gives credit to Segal saying, “He’s the one who wound us up and turned us loose. He’s the one that lit the fuse with us.”

Following his conversation with Segal, St. John teamed up with his morning weightlifting buddies, Craig Eisler and Eric Engstrom, who, according to St. John, “had nearly finished crippling Borland in the Windows tools business and were also looking at new opportunities,” and began to conspire to change the world—again.

St. John and the Calculated Approach

St. John puts the two faces of the DRG in perspective. “Again, when I talk about DRG and tell the funny stories about it, because those are fun for everybody to read, underlying it was a very, very calculated approach to using a small handful of people with a particular world view to have enormous influence on a market. When you think about the classic picture of a lever, DRG was one of the most leveraged forces that Microsoft ever contrived to influence markets. Historically, that’s so important to understand about it, because people who never lived or experienced that, or who were outside of Microsoft, it’s a constant mystery why platform companies like Microsoft are hugely successful. ‘They’re lucky.’ ‘I don’t get it.’ ‘It’s unfair.’ Or, ‘I think they’re cheating somehow, but I can’t figure out how.’ And I know one aspect of it. You know. Just competing. Just fighting hard. Working the press and working the developers and platforms to make sure that you come out on top. To put it bluntly, we were monstrously good at it. And in that context, this was not a group whose job it was to design platforms. It was our job to sell the crap that Microsoft was making, and I call it crap because a lot of it was. And that’s one of the reasons that’s so interesting about the DirectX story, is that James taught us that our job is to get mindshare. Our job is to get people adopting and tangled in Microsoft APIs.

“When I looked back, I said wow. I was so young and dumb, I had no idea what I was a part of, or fully appreciating the scale of it. I was just in there going, ‘Yeah. This is great. Whatever. I’m going to win.’

Nathan Myrhvold: the Semantic Level

With DOS about to disappear, what was the strategy going forward? Microsoft Research founder Nathan Myhrvold talks about the challenges for the game development community.

“There is always an issue in battling for attention; do you go for the next little version of the incremental thing, or do you go for something that’s bigger? And there’s no answer to that. There are cases where people try to create the grandiose system that doesn’t actually work yet, and then there’s also plenty of cases where they make the other mistake by going with what seems to be tried and true and immediate, and that misses the bigger picture. One of the issues, for example, with games, is that game companies and game developers in a certain era were the last of the hackers who wanted to do everything themselves. And it was a strategy that was completely destined to fail, but it was very strongly felt by them. And what I mean by destined to fail is, as you get more and more capable graphics chips, and as you start targeting different platforms, you need to raise the semantic level at which you program with the system. But the earliest games were written in assembly language by people trying to get every tiny gram of performance, not that performance is measured in grams, but you know what I mean. They were trying to do absolutely everything possible, so they had an attitude about what was the right thing to do. And all of the different approaches, whether it was OpenGL or DirectX or other things, were all different approaches to try to raise the semantic level of the interface to graphics. Because prior to that, all of the graphics stuff had been done by games very, very directly.”