The assembled visitors dutifully laughed at Taschek’s oblique reference to the growing rivalry between Microsoft and the small companies pushing Linux as an alternative to Windows, but the frost in the air didn’t dissipate one bit. Grouchiness was the order of the day. Excluding the journalists, no one really appeared to be enjoying themselves. The tour itself was a distraction from the main event: the commencement of a week of intensive comparison testing between the Windows NT and Linux-based operating systems.
The representatives of the Linux community — two engineers from Red Hat, the leading vendor of packaged Linux distributions, and one from Penguin Computing, a fast-growing seller of computer systems running Linux — made no attempt to sugar-coat their feeling that the tests were unfair and irrelevant, that the particular configuration of hardware chosen favored Microsoft. For their part, the delegates from Redmond, led by Jim Ewel, a veteran Microsoft marketing specialist recently fingered by the Wall Street Journal as the leader of a newly formed “Linux attack team,” responded to the obvious disdain broadcast by the Linux coders with their own imperturable brand of self-confidence. If the Linux punks wanted to make good on their claims to be aiming at the big-time “enterprise” market, then they had to learn how to play with the big boys.
Even Bruce Weiner looked uncomfortable, although he appeared confident that the PCWeek tests will ultimately ratify the controversial results his company trumpeted in a study released last April. That report declared NT to be far superior to Linux — and incited a storm of outrage from Linux fans, who were incensed when they learned that the test had been commissioned by Microsoft and was conducted in a Microsoft performance testing laboratory with the help of Microsoft employees.
Uncomfortable, strained, replete with icy glares — this was not your typical industry get-together, with its hypocritical politeness and warmed-over marketing hype. Instead, it was an initial skirmish between determined opponents from opposite ends of the computing-culture spectrum, circling each other like growling dogs trying to decide whether to fight.
The oddest discontinuity was to see Microsoft sitting across the table from Red Hat — as if the two companies were equals of some sort. Never mind the absurd financial disparity between Red Hat — a company with $10 million in revenue that spent $2 million on research and development in its last fiscal year — and Microsoft, which raked in $14.5 billion and spent a whopping $2 billion on R&D. The cultural differences were even more significant
Red Hat and Penguin sent coders to the table — young, confident programmers who get their kicks hacking on the Linux kernel, the heart of code that serves as the engine of the Linux-based operating system. Microsoft sent a team of marketers — a public relations account executive from Waggener-Edstrom, one of Microsoft’s stable of PR firms; a product manager for Windows NT; and Ewel, a director of marketing who has done stints as a point man for Microsoft’s SQL database efforts and Windows 2000 campaign.
Marketers and geeks — never the twain shall meet. But even so, the sight of Red Hat and Microsoft business cards mixing on the table offered undeniable proof that the established software marketplace is long past the day when it could blithely ignore open-source software — or deride it as “freeware” worth the price you paid for it. A year ago, reporters seeking comments from Microsoft executives about Linux were stonewalled; today, Microsoft is taking Linux seriously.
“We do see them as a competitor,” acknowledged Ewel.
Early on in the portion of the meeting attended by reporters, the conversation centered on how to interpret the numbers that would be generated by the performance tests. Such numbers are known as benchmarks — snapshots that capture how a particular software program performs on a particular hardware configuration. To the Linux camp, the benchmark process — or “bench-marketing,” as Penguin’s Mark Willey derisively referred to it — is fundamentally flawed. Benchmarks don’t reflect real-world situations, and are too easy to manipulate.
But to the Microsoft camp, as well as to Bruce Weiner, such carping about the irrelevance of benchmarks sounded like the venting of sore losers. Repeatedly, Jim Ewel made the case that purchasing managers at big companies require hard data to justify orders that could amount to millions of dollars. In that arena, benchmarks serve as a crucial component in plotting serious corporate computing strategy.
Weiner bridled at the unvarnished contempt that the Linux camp expressed for Mindcraft’s original tests. After the second instance in which Willey used his fingers to mime quotation marks around the words “independent test lab” — implying that Mindcraft was nothing more than a willing Microsoft pawn — Weiner leaned across the table and told Willey, “You are challenging my integrity.”
With Willey, young and brash, facing off against the middle-aged Weiner,
it wasn’t hard to see that little interchange as symbolic of a much larger generational confrontation in the software world. But it wasn’t the only revelatory moment.
I asked the Linux programmers whether the hype that had accompanied the Mindcraft tests had affected the rate of kernel development. After all, one of the key strengths of the open-source software development model is supposed to be the flexibility and speed with which it responds to newly discovered bugs and operating flaws. Zach Brown shrugged, and said a couple of programmers had spent a weekend or two hacking together solutions to the main problems demonstrated by the Mindcraft test. The changes to the kernel weren’t yet stable enough to be employed in the current PCWeek tests, said Brown, but as far as he was concerned, the Linux development process had already moved along — it hadn’t been accelerated by Mindcraft, it had merely chugged along at its normal blistering pace.
Weiner and Ewel seemed a little thrown by Brown’s blithe arrogance. If there’s one thing that the open-source programmers can definitely match Microsoft at, it’s in their absolute certainty, at least in public, of their own unassailable rightness. No doubts, no hesitation — the Linux programmers didn’t appear to really care that the numbers that will result from the PCWeek Lab tests will probably show Windows NT in a positive light. That’s the present, soon to be the past. The future is theirs.
Is such arrogance prudent when dealing with an opponent like Microsoft? You certainly couldn’t imagine a better way to get Microsoft’s direct attention than to strut around with such an attitude.
But just what does Microsoft think about open source? Even in the loose atmosphere of the PCWeek Labs conference room, it was difficult to get Ewel to speak in more than just the same old tired generalizations that Microsoft officials always make in their public comments.
It’s all about the customer, said Ewel. If the customers want source code, Microsoft will consider giving them access to it. Microsoft customers had already demonstrated a distinct lack of interest in having a version of Microsoft Office that worked on Linux, so Microsoft had no plans to push forward in that market. As for open source’s vaunted speed and flexibility? Now it was Ewel’s turn to shrug. Internet Explorer 5.0 is way ahead of Netscape’s open source browser project, Mozilla, said Ewel. All those daily updates to the Linux kernel? The NT kernel, said Ewel, is also improved every day, recompiled every day, and tested every day.
But ultimately, Ewel stated, Microsoft is an “intellectual property” company.
“We believe that people getting paid for their software works. We do strongly disagree with GPL licenses,” he said, referring to the Free Software Foundation’s GNU general public license, which governs much open-source software and mandates that the source code to a software program must be accessible to all and freely modifiable by all.
And on that note, the press-attended portion of this latest installment in the NT-Linux shoot-out ended. The programmers strolled over the to lab, where the racks of Pentium 233-powered computers stood ready to hammer upon the chosen server. A week of tweaks and tunes will follow, and no doubt there will be some serious geek bickering before its all over. However well the Linux hackers perform, when PCWeek Labs releases the results in at least another week, you probably wouldn’t do badly placing your bets on Microsoft, this round. But we could be in for a much longer match.
Does it have to be? Every now and then, a strain of thought surfaces in the debate about open source that suggests that the dynamic between Linux and Microsoft doesn’t have to be an either-or, zero-sum equation — that there is plenty of room in the world for multiple operating systems. The hackers can have their Linux, while corporations will play it safe with Windows.
But that line of thinking underestimates both Microsoft’s paranoia and open-source ambition. It underestimates how every Microsoft attempt to cast Linux in a negative light will inspire open-source programmers to redouble their efforts. The “intellectual property company” with its billions of dollars and the international community of free software programmers will undoubtedly continue to clash — and send more sparks through the software marketplace. Both sides think they know best — a state of affairs that usually precedes a knock-down, drag-out fight.
Forget the antitrust trial. Ladies and gentlemen, we have a competition.