Tim Bray at JavaOne:

After saying some nice things about Rich, Jonathan proposed that they do a Q&A, with Jonathan asking the questions, saying a “I’ll simulate a developer”. His first question was “So Rich, are you going to open-source Java?” Rich started with “Well, why not?”

Here's a "why not" for Sun to ponder: If you OSS Java (and never mind the compatibility issues Tim mentions later in the post), then the license revenue Java does bring in dries up. Sun already destroyed the value of Solaris by open sourcing that - by open sourcing Java, the (admittedly paltry) revenues from that will dry up and blow away.

At that point, someone is going to have to go into the exec offices and ask exactly what the value of Java is to Sun. It speeded the commoditization of Solaris and SPARC - once money makers for Sun. The guy who takes over Sun after Schwartz drives it into the ground is going to have very little of value left to sell - either to customers or parties interested in acquiring the remnants. As Forbes says:

What pieces could Sun sell? The three most valuable assets are the Sparc chip line, the Solaris software and the hardware that uses both. The three would be hard to sell independently because they're designed to be tightly integrated. But the chips, once the clear choice for high-performance computing, are no longer on the cutting edge. And Sun cut the value of Solaris steeply by turning it into an open-source operating system. StorageTek, a tape-drive maker Sun acquired last year for $3 billion net of cash, could be sold, but there isn't exactly a line going out the door of people who want to buy tape-drive companies. Sun's Java software language generates lots of goodwill among programmers but not much revenue. Kariithi estimates it brings in only $10 million a year.

Somewhere, the folks at IBM who actually do make money on Java are laughing, a lot.