Is open sourcing Java a good idea for Sun? I think the answer is an unequivocal yes. I also think that the chances that Sun will actually open source it, especially since IBM suggested it is some where between slim and none. Darn it.
I mean, come on, Sun has been flipping and flopping about Eclipse, an open-source integrated development environment (IDE) which started out with IBM sponsorship, for ages now. And, lets face it. For all the words spent about Eclipse, an IDE isnt much of a deal compared to Java.
From where I sit, Sun has always been of two minds about open source. One part loves it, the other hates it. You can see the first side in Suns recent embrace of Linux on AMD processors. On the other hand, Sun embraced Linux once before, when it bought Cobalt Networks in a stock deal worth about a cool two billion dollars in 2000. Three years later, Sun closed the last of its Cobalt based network appliance line down.
I could go on with numerous other examples, but why bother, if you follow Sun at all you already know about their schizophrenic relationship with open source. IBM, on the other hand, has embraced open source.
Now, IBM doesnt support open source and Linux because of some starry-eyed idealism. IBM supports open source because its good programming and, far more importantly, its good business.
IBM realized that when you open source a program, you not only get a better program, you spread the software development cost risk. If IBM spends a hundred million on a program and it tanks, thats a hundred big ones down the drain. If IBM spends twenty million on an open-source project, Oracle contributes ten million, Apache shells out another five million and a programmer spends some $100-an-hour-time on the project and it all adds up to a hundred million of work, IBMs financial exposure is a lot lower.
Theres a reason why IBM never became a Linux distributor. IBM realized that it didnt have to spend much on Linux, compared to creating its own proprietary operating system or own it, to reap the benefits of both lower financial risk and higher service revenues.
Suns problem, as I see it, is that it has never made that jump. Deep in its heart, Sun believes that the only way itll see real profits is by selling SPARC boxes that run Solaris. You only have to look at IBM and Suns respective financial track records to see which approach is working out.
This isnt to say that IBM simply leeches off open source. They contribute quite a lot both directly via alphaWorks and indirectly by paying for Linux and other open-source development. Sun, on the other hand, while it has some open-source projects, Project JXTA for peer-to-peer networking and OpenOffice, keeps its distance from going whole-hog into open source. Thats why we have a StarOffice.