A few years ago, I pretty much wrote off Sun Microsystems as a technological has-been. SPARC processors were falling behind in speed compared to Intel and appeared to be headed to the same dust heap as MIPS and PA-RISC. Solaris was losing ground to upstart Linux, while all Unix vendors fought for a piece of an ever-shrinking pie. The NetBeans IDE was so slow that it was used as an example of why large apps could not be written in Java. In fact the only thing that Sun appeared to have of value was Java itself, Imagine if Solaris had been open sourced before Linux gained a foothold. but it was held so closely that it was almost smothered by its creators.
What a difference a few years made. NetBeans turned a corner and started being, well, good. Sun open sourced Solaris and came out with Solaris 10 containing may innovative features like dtrace. They came out with - gasp - price competitive workstations and servers based on AMD processors. And on the SPARC side, rather than let it die they came out with a clever multi-threaded version called Niagara that just might turn conventional wisdom on how to build database servers on its head. Suddenly Sun is charging ahead with all barrels blazing. It reminds me a bit of how IBM re-invented itself and kept itself from becoming the dinosaur everyone thought they were. How did that happen?
There are many theories but I think the main reason is that Sun has embraced open source almost all the way through their organization. They've even 'open sourced' their hardware architecture. You could argue that open source helped rescue IBM too, but until IBM open sources AIX, zOS, and WebSphere, there is no way that IBM has embraced it to the extent that Sun has.
If there's a down side it's that it took Sun so long to see the light. I'm not just talking about 'support for open source', or 'number of OSS projects', I'm talking about in their core, strategic areas. Imagine if Solaris had been open sourced before Linux gained a foothold, for example. Linux might have faded into obscurity instead of becoming the market force that it is today.
Of course Sun is still dragging their feet on Java, saying it's already open source (it's not) or that it doesn't matter that it's not open source (it does). Will they wait until Apache Harmony gains critical mass and surpasses Sun's Java in users and mindshare? History says yes, which is really too bad. They've gotten it right in so many other areas.