Well, they did it. For months we've been hearing about Sun Microsystems' plans to open the source code of its flagship Solaris operating system. Last week, the company finally delivered the goods. The question is: Will Sun be any better off?

Just five years ago, the idea that anyone but Sun's own engineers would be allowed access to Solaris internals seemed absurd. Back then, Sun was flying high. A Web start-up wouldn't have been taken seriously if its production sites weren't deployed on multiprocessor SPARC servers. Run a high-volume Oracle (Profile, Products, Articles) database on anything less than Solaris? Never.

So why would Sun stoop to this seeming indignity now?

Obviously, you don't need to be a 35-year-old programmer from Helsinki, Finland, to answer that question. Open source hit Sun like a ton of bricks. Slowly but surely, away went the purple Solaris boxes as companies such as E*Trade and Google (Profile, Products, Articles) passed up big iron in favour of low-cost, flexible Linux servers.

Today, both the Oracle and IBM (Profile, Products, Articles) DB2 databases list Linux as their preferred platform. PeopleSoft (Profile, Products, Articles) runs on Linux. So does SAP (Profile, Products, Articles). And Novell (Profile, Products, Articles) is using Linux as the basis of its entire suite of networking and identity management products -- a market Sun sorely covets.

All this must have Sun execs slapping their foreheads: "But Solaris is better!" And you know what? They're right. Pound for pound, there's probably more technical innovation in Solaris than any other Unix platform. The problem is that Sun, like Microsoft, didn't grasp the strengths of Linux until very late in the game. In Sun's case, maybe too late.

As the premier Unix vendor, Sun has long taken the attitude that Linux is nothing more than a hobbyist's clone of Unix. Why go with the copy when you can get the real thing? But what Sun failed to realize is that the rapid development pace of the Linux kernel meant that, sooner or later, Linux would inevitably leave the ranks of the amateurs and become a real competitor.

Sun's second mistake was in underestimating the appeal of a Unix OS that ran on Intel (Profile, Products, Articles) x86 processors. It's true that Sun servers are superior to commodity x86 hardware, but maybe not enough to justify their price tags. When Sun tried to abandon Solaris x86 in 2002, it left budget-conscious customers to select a Unix OS that didn't require expensive SPARC hardware.

OK, Sun fixed that mistake. Solaris x86 is back, and with it a new marketing push aimed at addressing the cost issue. Not only is Solaris price competitive, Sun says, but when all is said and done its total cost of ownership is actually less than that of Red Hat Linux. This is fine, except that everybody knows Red Hat isn't your only option for Linux support, and it's certainly not your only source for Linux.

And so, at last, Sun throws up its hands: "Fine. You want open source? Solaris is open source." And with that, away goes the last argument why anyone would choose Linux over Solaris -- at least in theory.

Unfortunately, in the time that it has taken Sun to figure things out, the tables have turned. Linux's feature list and its popularity continue to grow. For Solaris, some fancy features, a lower price tag, and open code won't be enough. To remain viable, it needs to grow its customer base. And this time, it's Solaris that needs to win over Linux customers, not the other way around.