Scott McNealy is stepping down as CEO of Sun. He will stay on as chairman.

Once a silicon darling, Sun has fallen on hard times since the dot-com bust. Their problems weren’t dot-com related, though; they’ve been torpedoed by some very basic business problems.

Their bread and butter was proprietary systems running their own version of Unix called Solaris (oddly named, given that the book and movie entity named Solaris drives people insane). Solaris was, in its heyday, easily the best such animal available. Sun’s hardware, too, was impressive at the time, and for a while at least was synonymous with computing power in a time when Intel hardware was barely capable of running Win3.1. We remember being amused and amazed when, in 1997, we helped deploy one of the first web stores with Oracle, Netscape Application Server, and the Netscape web server all on one fairly low-end Sun (an Ultra 5 — just like the one we use a doorstop in Heathen World HQ now). Of course, now the much-cheaper Xserve that runs Heathen is completely capable of running all that and more, and in fact does.

The the world changed. Intel hardware got much better, and did so very quickly. Linux came out of nowhere and caught all the expensive, proprietary Unixes flat-footed; the combination has been huge for Sun. Other firms were caught, too, but some managed to be more agile; IBM in particular adopted Linux as its own, and has been busily giving back to the FOSS community. Sun, on the other hand, became convinced that they could give away hardware and make it up on software — i.e., the still proprietary Solaris — and survive. They were, and are, wrong. Sun hardware may still be an acceptable choice for HUGE database applications, but those jobs are few and far between, and Sun no longer owns the market. For 95% of the applications, there’s no reason to buy fancy hardware or software, and that’s the problem. For 100% of the applications, there’s no reason to use Solaris and be tied to Sun when choosing a supported enterprise Linux keeps your options open.

Sun has given the world Java, of course. Their vision for Java has never been completely in line with reality, but that’s Silicon Valley for you. Java remains a solid contender for server-side development (and even the occasional client-side app); aspects of the language make it a much better choice than almost any other compiled language. However, you still need a pretty big app to justify it over options like the stack, or Ruby on Rails, or other open source offerings. The world of Java is often an overly complex one full of cargo-cult dogma and huge codebases nobody actually understands; more recent approaches tend toward the simple, with attendant increases in maintainability. Java was and is free, however, so it’s not about to fix Sun’s bottom line issues; the minute they try to charge for it is the minute the FOSS options bury it.

So that’s what Sun’s got: expensive hardware of dubious value, a white-elephant operating system, and a language/technology they gave away for free. Vanishingly few people want the first two, and the third can’t save them. And all this has been true for years. The surprise isn’t McNealy’s departure; the surprise is that it took the board this long to do it.