GAME OF X V.2

Windows continued to improve with each new release and to gain traction. Windows 286 and Windows 386 were increasingly successful and by 1988, Microsoft had become the world’s largest PC software company based on sales. Although adoption of the Windows platform had increased, it wasn’t universal, and so Microsoft promoted bundling the installation software for Windows in applications like PageMaker, which helped disseminate the OS into more and more offices and homes.

Finally, with Windows 3.0 and 3.1, Windows began to hit its real stride, and Microsoft even showed a rare sense of humor when they added Solitaire, Hearts and Minesweeper to the basic configuration and released an ad that stated, “Now you can use the incredible power of Windows 3.0 to goof off.” (There had been some similar humor in the ads for the entertainment packs, such as, “No more boring coffee breaks” and “You’ll never get out of the office,” but the Windows 3.0 ad was a very rare acknowledgment of games by the more serious OS and business software group.)

Windows 3.1 also shipped with Watson, a tool to help developers with debugging, and the evangelists of the Developer Relations Group were there to help developers learn to use the new tool. According to Myhrvold, “Watson added much better support from tools vendors. And that was also a major focus. We had a guy in my group named Bob Taniguchi who worked with all the tools vendors, and that was super important.”

So part of the DRG’s mission was being accomplished. The Windows platform was beginning to gain traction, and Microsoft’s technical evangelists were working hard to encourage developers to write Windows applications.

While hiring technical evangelists was one important way in which Microsoft’s developer evangelism differed from Apple’s, perhaps even more significantly, the DRG mission was always about making Microsoft’s platforms stronger and its competitors weaker. Everything they did was to further those two missions, and they found ways to do both at the same time. To do so, Myhrvold required people with very special qualities.

One of the most influential and outspoken, but rarely celebrated, Microsoft evangelists originally came from the Apple camp. James Plamondon had spent much of his youth in the wilderness of the redwood forests of Del Norte, California, just south of the Oregon border. His conservative Republican parents had moved out of Southern California, according to Plamondon, because they “were concerned that there was way too much sex and drugs going on in the local school, and so they wanted to get us out of that evil big city environment and back to nature.” Plamondon quickly points out the irony of that move because, with the timber industry and commercial fishing both collapsing in the north woods, “all that was left was growing dope.”

Because he had contracted polio as a child, Plamondon’s father was unable to join his college mates in the Korean War. Out of frustration, he became obsessed with military campaigns. “He had been following those wars with maps on the walls and listening to the radio reports… just following them as closely as he could.” Plamondon and his brothers were included in their father’s growing interest in military strategy, but it was their father’s other interest that most fascinated the Plamondon boys —computers. In 1977, Plamondon’s father brought a TRS-80 computer into the home, and even more than war history, the computer became young James’ obsession.

James Plamondon ultimately went to college and earned a degree in computer science with a focus on artificial intelligence. He was fascinated by the potential of using artificial intelligence in games. “And so I got this bachelor of computer science degree that was focused on artificial intelligence, and then—duh—after I did all that, I then did my market research, talking to various game companies, and they said, ‘You know there is absolutely no demand for what you’re offering. That’s just great and everything, but the reason it sucks is because we don’t care.’”

There was no plan B. It was time to shift gears. Fortunately Plamondon had gotten an early “phone book” edition of Inside Mac before the Macintosh was even released. He taught himself Mac programming, even though there weren’t any Macs to program—yet. When the Mac did come out, “I convinced my girlfriend at the time—who later became my wife—to buy a Macintosh so that she could start up her own print design business, which she actually did fairly well at for a while.” And, of course, Plamondon used his girlfriend’s Mac to hone his skills.

Soon Plamondon found work at Reference Software, a company started by semi-retired professor, Bruce Wampler, to market the world’s first spell-checker and thesaurus. After at one time selling his software to WordStar, Wampler was able to buy it back again when Word Perfect came to dominate the word processor market and WordStar began selling off its assets.

Plamondon worked for Wampler for about a year, while they sold the spell-checker/thesaurus software as an add-on for other products, but eventually he came to realize that the job was a dead end. “I’m the only other guy there. It’s him and me. What? Am I going to get promoted?”

Plamondon took other jobs, specifically choosing them for what they could teach him about programming. He had seen the promise of the graphical interface, and believed that it would reshape the world. On the other hand, there was DOS. “Mac was the vehicle of the graphic user interface, and was therefore morally superior to DOS. (spit) DOS (spit). You have to spit after saying DOS. Otherwise the taste of it might remain in your mouth and poison you later. I once wrote that DOS was proof that there was evil in the world.”

Completely convinced that the Mac was the future of computers, Plamondon began writing Mac programming articles.

Plamondon decided that he wanted to find jobs that would help him refine his Mac programming skills, and Silicon Valley was the place to find those jobs. So he moved. He got a job at a company that wrote statistics software, and it was there that he had a revelation that propelled him forward. “One of the guys, who was the senior programmer there, would write the same list management code over and over and over, so that he could rename the functions, so that when he was debugging he would know which list he was debugging because of the function names. Which meant that whenever there was a problem in his list management code, he’d have to change it N times, because there were N different versions of this code scattered around. He had different code for each different list. It was insane. So that led me to think there had to be a better way.” Which is how Plamondon discovered object oriented programming.

He found a job where he could learn more about this new kind of programming at a company called Power Up Software, which provided him with the opportunity to work with the first widely used object oriented application, MacApp. While he was working at Power Up Software, Plamondon organized an object oriented programming Framework group, which met at Apple’s headquarters. “I was already a mover and shaker in the Framework world,” he says, which led to an offer from Microsoft to write a book about their Microsoft Foundation classes. However, he was “stony cold broke” at the time, and had to turn down the offer, knowing that he wouldn’t see any money from the book for months, at best.

In the market for another job, Plamondon interviewed at Apple and at Taligent—a cooperative venture between Apple and IBM to build a new operating system. Plamondon described Taligent as “a real sheep lying down with the lions kind of thing” because Apple and IBM had been enemies, at least until Microsoft became their common rival.

Plamondon considered working at Taligent, even though he was sure the venture would fail. “They tried to write an object oriented operating system using C++, and C++ just wasn’t up to the task… It was like you were using a screwdriver to solve a hammer problem. And they just weren’t likely to succeed.” Still, he considered it, seeing it as an opportunity to learn valuable skills, getting out before it crashed and burned, thereby avoiding the “odor of failure” that would attach to those involved.

While he was musing the pros and cons of a job at Taligent, another opportunity presented itself. He was invited to interview for an evangelist job, not at Apple, but at the much-hated Microsoft. Why would he even consider the offer?

Cognitive Load

“It all comes down to cognitive load. With a graphic user interface, everything you need to know is embedded in the user interface. It may be four menus deep, which is bad user interface design. It may be a really shitty window… a bad dialog box or something, but it’s all right there. You don’t have to remember, and you don’t even have to remember how to get to it, if the interface is sufficiently explorable, so that you can discover how to use it. The ideal graphic user interface program has no manual because it is so self-explanatory. And I just loved that to pieces, because if you thought computing was a powerful and a good tool—those two things together: powerful and good, liberating creativity, which everybody of my generation, and probably yours, believed was true—then you wanted to empower as high a percentage of the population to use computers as possible. And the command line interface was a barrier to that because you had to fucking remember everything. It’s not discoverable. You have to memorize everything. Every secretary on the planet had this little Word Perfect cheat sheet on the top of the first drawer to her left. Right? Something where she could pull it out and check, ‘How do I do this again?’ It made everybody feel stupid. And it was the same with DOS commands… So my goal in my early stage of computerdom, was to drive this whole graphic user interface notion better because it was a way of democratizing computer usage—empowering computer users.”

-James Plamondon

Why abandon Apple? Simple. According to Plamondon, Apple had abandoned him, and all the other small developers.

Plamondon’s personal goal—his mission—was to “empower users with computing power… to bring the power of computing to the masses.” As he saw it, Apple had lost track of their identity. In the early days of the Macintosh computer, it was “the computer for the rest of us,” but by the late ‘80s, they had changed. In Plamondon’s words, “They were now the Mercedes of computers.” They catered to the elite customers while people who used DOS and Windows were driving Fords and Chevys.

“Well, first of all, that’s antithetical to what they set out to do. It’s the exact opposite of what they were saying originally, and what I signed on for, and everybody else I knew who was writing Mac programs. Secondly, it was selling their developers down the river because if a third-party developer targets a platform that he thinks is a Chevy, it’s because a Chevy sells a shitload of copies, and therefore he has a big platform to sell into. But if suddenly he’s now selling into the Mercedes market, his installed base has just shrunk like crazy. His opportunity to sell copies has just collapsed. Therefore he must jack up his prices beyond all reason. And so the ability to make a living as a developer of third-party Mac applications collapsed in the late 80s, early 90s because Apple took this change of focus from the computer for the rest of us to we’re the elite computer. Really pissed me off. They just stabled all their developers—my friends.”

Moreover, Plamondon asserts that Apple did not truly appreciate its independent developers. He cites a story of one senior vice president at Apple calling all Mac developers parasites. Plamondon also tells a story about Steve Jobs and the Next machine (see next page).

Plamondon decided, what the heck, why not interview for Microsoft’s technical evangelist role? To him, the important quality that differentiated Microsoft’s evangelism from Apple’s was the same distinction that Myrhvold had made—it was technical. It was a role that fit him perfectly. Combining his love of code, his love of sharing information, and his amateur military studies, he came in prepared. In fact, Cameron Myhrvold hired him on the spot. “He’s a guy who walked into my office with a mission. I mean he walked in to the interview knowing exactly what he wanted to do. Those are sometimes the easiest interviews. It’s like, ‘Oh. Wow. OK. How would that work? Tell me about it.’ And by the end of the interview, you’re like, ‘Jesus Christ. Gotta hire this guy.’ I mean, he’s written the battle plan.”

At first, Plamondon remained in the Bay Area, now working for Microsoft at what they somewhat facetiously called the “Bay Area Embassy.” This was just before the release of Windows version 3.1.1. “The Bay Area viewed Microsoft as being the evil, dark enemy up in Seattle,” says Plamondon.

The Next Machine

“Steve Jobs, when he introduced the Next machine, he had an event called the Next Day; because it was after a presentation to financial analysts and stuff in New York. On the next day he came to Silicon Valley and pitched to software developers, and I was one of the developers he was pitching to in the audience. And everybody loved the Next operating system and the interface builder and all these wonderful object-oriented libraries. It was using Objective C, which was one of the early dynamic languages. You could do it in Objective C. You couldn’t in C++. And everybody was just, ‘Oh, this is such an exciting thing. I’m dying to do this.’ And then he said, ‘Oh, and by the way, all of your software will be delivered on this optical drive that we’ll send to every customer once every three months.’ And somebody raised their hand and said, ‘What? What? Did you just say that I can’t distribute software myself? That I have to distribute my software through you?’ And Steve Jobs said, ‘Nonononono. You’re thinking of this wrong. We’re not taking away from you the freedom to distribute your own software. We’re enabling you to not have to worry about software distribution at all. And we’ll only take a 30% cut.’ And we all stood up and walked out of the room. Almost literally. At that point, we all looked at each other and said, ‘Oh, after you.’ ‘No, after you.’ ‘Oh no, no. After you.’ And nobody wrote for the Next machine because nobody was willing to hand over all their distribution to Steve Jobs and Next. What’s funny is… He always wanted to get a piece of the developers’ action. He always felt, and Apple has always felt, that application developers were parasites, and that there had to be a way to get a piece of their action. And they finally figured it out with the App Store. With the App Store, they’re skimming the cream off the top. It doesn’t matter whether you sell one copy or a million copies. It’s like the patent business. Patent attorneys get their money up front, before you’ve made a dime selling your invention. It’s the same way with the App Store. They’re making their percent on everything that sells, whether it sells one copy or a million copies.”

-James Plamondon

Network Effect

According to Plamondon, one of Apple’s chief failures was their failure to understand the “network effect,” which he describes as “positive network externalities and the relationships between applications and platforms and tools.” According to the Financial Times, “A network effect (also known as network externality) exists when a product’s value to the user increases as the number of users of the product grows.”

Understanding the network effect is critical to promoting platforms, and if you’re in the platform business, you want to encourage as many applications as possible to use your platform. Having more applications increases the value of the platform to the user. And, according to Plamondon, you want to do all you can to help these applications to be successful, not out of altruism, but because it makes your platform stronger, and the stronger your platform becomes, the weaker rival platforms are by comparison.

Part of helping developers succeed is to develop tools internally, or to encourage and support others to develop tools. While tools are important to facilitate application development, they are a loss leader. “If you try to profit from the tools, you’re raising the cost of targeting your platform.”

Plamondon asserts that Apple never understood the network effect or the platform business. He even cites how Google’s Android mobile operating system has so quickly encroached on Apple’s seemingly dominant IOS business. Back in the late 1980s and early 1990s, Apple lost market share, at least in part, because they didn’t treat their Mac OS as a platform the way Microsoft did Windows.

At Microsoft, Plamondon found people who understood the network effect and the importance of empowering developers to make applications for your operating system, starting with Cameron Myhrvold. And Myhrvold also gives major props to Bill Gates and Steve Ballmer, commenting, “They completely believed, just in their bones, that that was true. When I would talk with other people who had my role in other companies, they’d have to go through these massive justification exercises. I never had to do that. These guys just knew.”

Myhrvold describes Apple as “the counter culture and the prima donna of the UI,” and although they were effective, he points out their critical weakness. They thought like a hardware vendor. “If you think like a hardware vendor, you look at a pie and you think, how is the pie going to be divided between hardware and software? And that is what causes you to then deliberately think about how to make software cheaper because it’s a zero sum game, and that you’re going to sell more hardware.” The problem is, once again, the network effect. Apple wanted to sell hardware, and they believed that cheaper software was the way to do that, but they didn’t seem to appreciate the importance of their independent developers the way Microsoft did, nor did they seem to understand how to use the developers as pawns in the operating system game of chess. Because they were trying to sell hardware, and Microsoft wasn’t.

Even though they were a developer, too, Microsoft understood that the more people developing for their platform—in this case, the Windows operating system—the more indispensable that platform became. Plamondon saw at Microsoft an opportunity to support developers and to proselytize his idea of computers for the masses. In his mind, Apple had become the bad guys, and Microsoft—at least the people in the DRG—were the good guys fighting the good fight.

In part, developers hated Microsoft in the early days of Windows because of their policy of consolidating Office applications into bundles, which made it harder for independent developers to compete. Of course bundling was purely a marketing decision. Describing the initial bundle of Word, Excel and PowerPoint, Plamondon observes, “They had different menu structures, and they had different core code. They had almost nothing in common. It was as if they were from separate companies.” Clearly a kluge, the bundle was a good deal due to Microsoft’s aggressive pricing. During the years following the launch of Office 1, Microsoft added Outlook to the basic package, and in some configurations, Access.

Office also became more consistent, with menu structures and commands being shared among the different applications, but that consistency did not come easily. At the time, different product divisions at Microsoft operated almost like small, independent businesses, what Plamondon called a keiretsu, a Japanese word describing a set of companies with interlocking business relationships and shareholdings.

Plamondon explains, “It was a Japanese keiretsu that basically just provided capital and a thin layer of high-level management to a bunch of individual companies that were product lines. And so the powerful people in the company were the PUMs, the Product Unit Managers. He pronounces it ‘pum’. And if you were the Excel PUM, you did whatever the fuck you wanted to with Excel, as long as you could keep driving its numbers up. You were the king of Excel. And if you were the PUM of Word, you were the king of Word. You did whatever was necessary to sell more copies of Word. And if it fucked Excel, that was too damned bad for Excel. And there was a thin layer of tax that was imposed on this. For example, I think it was the Word guys came up with OLE (Object Linking and Embedding) because they wanted to be able to bring in Excel objects—particularly Excel, but from other things as well—and manage them from within Word. And the Excel guys didn’t want to put Word objects into Excel’s spreadsheets, so they didn’t care. But the Word guys cared a lot, so they created this OLE and component object model infrastructure to make that happen. And it was kluged up. It was a grotesque kluge, OLE 1.0, but they fixed the technology in OLE 2.0 by creating COM, the Component Object Model.”

Plamondon’s job description—the public version—was to bring Mac developers to Windows, and because he was known as a “Mac guy,” he had a certain advantage. “For at least a couple of years I was able to say, ‘I’m not saying X because I work for Microsoft. I work for Microsoft because X is true. X is cool. X is a pretty compelling technology. And you should consider using X because it’s awesome. And I’m not just saying that because I work for Microsoft. I work for Microsoft because they’re producing cool technology.’ But after a while, after you’ve been working for Microsoft for a couple of years, even if that’s true, it’s not credible anymore.”

COM: The Common Object Model

About the Common Object Model, DRG operative Alex St. John explains what it is, what it was meant to do, why developers hated it, and why it was necessary: “Common Object Model. It’s a technology Microsoft cooked up to enable API’s to change and evolve overtime while remaining backwards compatible. For example DirectX could evolve frequently as it did without breaking earlier applications that used it because of COM. Non-COM based API’s are just frozen in time for years in-order to preserve stability of all of the software that uses them. COM was clever but also very cumbersome for developers to adopt.

“It started out rough and got better fast. It’s major sin was that we felt that we had to support COM which was a big push at msft back then and made API’s much more complicated. OpenGL being an open-standard API couldn’t be “COMified” and be OpenGL. So there was certainly a lot of developer rejection of COM. We hated doing it. The other big issue with developers was the execute buffer API. You couldn’t just draw a triangle in early D3D, you had to fill out a really complex data structure called an execute buffer. The reason for it at the time was that 3D consumer hardware acceleration was very limited and using execute buffers helped them get acceleration. Without using them, a lot of hardware didn’t actually accelerate anything. The operating system couldn’t automate filling out an execute buffer for a game without losing performance because only the game knew how many polygons it intended to draw. It was a pain and confusing for the developers. Being pragmatic our guys said, why do the work to add an API that goes slower?

“It was stuff like that that worked itself out with time. In that example OpenGL had a draw triangle API that would simply have run slow and then the developers would have complained about that. It was the nature of developer relations.”

Guy Kawasaki had a degree in psychology and later obtained an MBA. He wasn’t a programmer. He didn’t code, and he didn’t talk code. Microsoft’s evangelists were all—every one of them—technical people who could sit down with developers and discuss their technical problems and needs meaningfully.

Plamondon describes the contrast, saying “Aristotle wrote a book called ‘On Rhetoric,’ and everything he wrote about rhetoric, about persuading audiences, is absolutely as true today as it was then. The only thing that he missed is the concept of identification—that you need to get your audience to identify with you, to think that you are one of them—because the only people he could imagine speaking to were fellow Greeks. Who’d want to convince a barbarian of anything? They’re barbarians. You don’t want their support.

“So this notion that you get people to identify with you, he didn’t get. And it’s really critical to what we did. We always made sure that when we were going to talk to doctors—technologists in the medical field—we’d send a doctor, somebody who was a medical professional who we’d hired because he was also keen on computers. And Microsoft in the ‘90s was able to hire those guys. So… empowerment. It’s all down to empowerment. So we at Microsoft believed that we were the good guys empowering the masses. Apple was selling this elite computer to the elite artsy fartsy people who we hated in high school anyway, and we were kamikaze nerds, fighting on behalf of secretaries everywhere, to make their lives easier, to make it so that our grandmothers could use computers, to bring computing power to the masses, a role that Apple had originally taken… and had abdicated.”

Plamondon on MortWare

“My favorite example is a company called MortWare… MortWare: the Software for Morticians. And if you look up on YouTube, I think you’ll find a video in which I talk about why the Developer Relations Group was able to kick the rest of the world’s ass. It was because we empowered some young kid. His parents ran a mortician’s thing, and he really wasn’t into the mortician thing, but he really liked computers. He could get started with Visual Basic and write an application that would enable his parents’ funeral parlor to run better. That would satisfy his desire to improve the operations, and to be empowered. All the things we wanted. He could do that without a computer science degree… without anything like that. To try to do that on UNIX, he would have to be a UNIX guru. He’d have to have at least four years of school. You know the tools were obscure. They’re all command line stuff’ that you have to have a PhD to figure out how to read the manual, let alone how to use the software. We took that guy by the hand and brought him into programming in little baby steps and made it really easy so that he could help his family’s undertaking business.”

Being the good guys wasn’t necessarily a matter of fighting evil. Their main rivals at the time—IBM and Apple—were not necessarily the real threats to Microsoft’s dominance of the personal computer space. IBM had long ago lost interest in the personal computer space, and Apple was on the decline. The ongoing challenge was to bring new people to the market, as Plamondon put it, “creating a solution for a problem that people had never been able to solve before.” New solutions included bringing computers into retail stores, doctors’ offices, auto shops… and even mortuaries. Their goal—Gates’ goal—was a computer on every desk and every home, running Microsoft software

The DRG, as Myhrvold saw it, was a bit more complex than Apple’s version, and when he first met James Plamondon, he recognized that here was a man with a plan. Plamondon already understood why Apple was failing to live up to their early promise, and he also saw that Microsoft had the potential to usher in an age of computer empowerment, and to accomplish that goal, he had a take no prisoners attitude. This was a campaign, much like those war maps he had studied with his father as a child. And there were rules of engagement, and if you followed those rules, you would prevail.

“Cameron Myhrvold is a brilliant guy,” says Plamondon, “and he understood network effects extremely well. And so he understood what it was that the group needed to get done. And his degree was in rhetoric… not kidding. So he was a professional persuader. He understood what needed to be done, and wasn’t exactly sure how to do it, that was clear. By this I mean no criticism. Nobody else did either. And I decided, since no one knew what to do—I didn’t know what to do, either—that I’d try something and see if it worked. And what I decided to apply to it was military strategy and tactics, because at least there was a well-established body of strategy and tactics there. And I happened to have always been interested in it. It’s not like I went to West Point or anything, but I was a dilettante in these things, had read The Art of War and various other books of military strategy.”

So while the evangelists of the Developer Relations Group were busily reaching out to developers and helping them solve problems and genuinely rooting for them to succeed (for admittedly selfish reasons), internally, Plamondon was developing a battle plan—not for any particular battle, but for each and every battle still to come.