Tim,

I can't help thinking there's a lot of simple-minded analysis out there regarding Microsoft's latest comments that Linux is the top threat to Windows.

I think that this is a VERY transparent attempt by Microsoft to get out of their anti-trust lawsuit through the back door. What a coincidence that they made the statement days before a Republican president, who has said he will drop the lawsuit, takes office. "Oh, we really believe that big bad Linux is a threat to our monopoly." I've always thought that this lawsuit was full of maneuvering on both sides: AOL buying Netscape but not shipping it in their product because they're locked in to IE (so they claim) in order to strengthen the case; Compaq claiming that they're locked into using IE even though they'd prefer to use Netscape (Why? That claim was made when NS was at v4); etc.

I don't believe that Microsoft will ever realistically compete with Linux or UNIX on the server. Not just because I think they have a bad product, but because they've shot themselves in the foot so many times by releasing such an inferior operating system. People will never trust them. I'm grudgingly impressed by the relative stability of Win2K, but I wouldn't entrust my website to Microsoft. It is beyond me why they don't take a stable core and build their pretty UI on top of it for their servers.

They clearly own the desktop, but is Linux even a threat there? Are they really so stupid to believe that they can compete on the server?

You've probably heard all this before, but all the recent hoopla about Microsoft's "earth-shattering" statement is frustrating.

Matt

I tend to agree with you that Microsoft's statements about the threat from Linux (and particularly their timing) are likely a tactic in the antitrust lawsuit. After all, as a desktop operating system, Linux has about the same share as the Macintosh. On the server side, though, Linux has injected new momentum into the UNIX market. As I've pointed out previously, Linux and UNIX together do constitute the leading server operating system. While the combined market share is only a little ahead of that for Microsoft operating systems, it's pretty clear to me at least that the momentum has shifted back to UNIX/Linux.

And as it turns out, momentum is very important in software adoption. Until Linux took off, it was becoming a self-fulfilling prophecy that the days of UNIX were numbered. Now, the sense that Microsoft success with NT (and now Windows 2000) was inevitable has been replaced by the feeling that Microsoft just might not make it to the top this time.

But at a deeper level, this posturing (on both sides) hides a deeper issue. Operating systems just don't matter the way they used to. The press tends to focus on the marketplace of the past, not the marketplace of the future. But the days of Microsoft's dominance are ending not because of Linux but because the Web has become the new platform. Netscape made this bold assertion in 1995, and Microsoft set out (successfully) to crush them. But while they could crush the leading browser vendor, they couldn't crush the Web itself.

In particular, in the years since Bill Gates announced that it was Apache rather than Netscape that was its biggest competitor for the Web, Microsoft hasn't been able to make a dent in Apache's market share. Of course, Apache doesn't stand alone: it runs on a Linux or UNIX (commercial or BSD) operating system, often with an open source database backend such as MySQL and with programming in open source languages like Perl, Python, or PHP. (For this reason, a lot of us are starting to use the acronym LAMP (Linux-Apache-MySQL-Perl|Python|PHP) to describe the open source Web platform. In the past seven years, LAMP has become the foundation for many of the Web's most successful businesses.

That being said, I don't at all agree with you that Windows 2000 can't compete on the server. While IIS is a clear second to Apache in the Web server market, it still has a pretty respectable percentage. And in e-commerce Web sites, I believe that they have the lead, or close to it. There isn't a lot of open source middleware; projects like Enhydra seek to fix that. However, partly because of the religious refusal of many Linux zealots to embrace Java, Enhydra hasn't gained as much visibility or traction as it probably should have. I'd really like to see a lot more common cause between the open source and Java communities.

Windows 2000 is, by all accounts, Microsoft's most stable server offering yet. And I'm fairly certain that they will build on this foundation, and will make an increasingly creditable offering in the server space. Microsoft is incredibly persistent, and really good at business development and building markets. But that's only the beginning. The Web is morphing into a platform for a new generation of network-based services. Microsoft's .NET strategy is incredibly far-reaching, and way ahead of the thinking in most open source companies, which have been fixated on the old battleground (the desktop) rather than extending the lead that they had in the new network-centric platform.

Meanwhile, Sun has been articulating a network-centric vision of computing since long before Microsoft got the new religion. But as noted above, the Linux community has missed a big opportunity to extend its offerings by embracing Java. (Fortunately, the Apache community has been more open-minded, and Jakarta includes a lot of interesting Java-related projects.)

At any rate, this story isn't over yet. Fortunately, we're at a new inflection point. The game is wide open, and Microsoft, Sun, and the open source communities all have key strategic advantages. The real question is who will best execute on the opportunity to take the emerging network-centric operating system to the next level.

--Tim

Copyright © 2007 O'Reilly Media, Inc.