With Sun releasing OpenJDK into the Free Software community at JavaOne three days ago, the landscape of Free Java will shift.
Stacy Cowley, CRN: One high point of the session came when Red Hat engineer Tom Tromey essentially offered to throw the project he's worked on for nearly a decade, Java compiler GCJ, onto the funeral pyre.
Panelist Dalibor Topic also sees a dimming future for a project he co-maintains called Kaffe, an open-source initiative to build a Java virtual machine (VM).
One panelist took a markedly different view, arguing that the Java community needs rivals to Sun's official implementation. Geir Magnusson, Jr., chair of the Apache Software Foundation's Harmony project, said Harmony still draws participants and plans to stick around.
I wrote an article about GCJ 1591 days (four years) ago. Looking back, I can clearly see that a major role it served is it steered Java toward the direction of Free Software. With the release of OpenJDK, GCJ and GNU Classpath (and other projects) should feel its historical mission fulfilled.
The fate of project Harmony is less certain. From a software ecosystem point of view, the existence of competing implementations tend to keep the technology moving forward. A Free Software license alone will not assure liveliness of OpenJDK. Even the venerable GCC needed a little fork (the egcs project) to push it out of stagnation in the 90's. So Apache Harmony staying around could only be a good thing for Java. And the interim governing board would do better to think long term, to a time when Java may no longer be strategic for Sun. (For example, what are they going to do after Java 7? Add a macro system?)
Incidentally, Per Bothner, who played a part in both the egcs project and the GCJ project, announce on the Kawa (a language implementation toolkit for the JVM, with Scheme and Xquery bytecode compilers, see my article here) mailing list three days ago:
Per Bothner: I'd like to let the Kawa community know that as of a week ago I became a Sun employee. I'll be working 80% time for Sun and 80% on Kawa (well, perhaps a little less). The hire process took a lot longer than expected because of the Kawa IP issues and resolving potential conflicts of interest with my existing Kawa support/consulting business. (No real disagreements; just figuring everything out.)
I'm working on the JavaFX project (which used to called F3), which was just announced this morning. See
My main task is compiling JavaFX to bytecode. There are obvious Kawa synergies and parallels; in fact I've been hacking together a prototype compiler for JavaFX using Kawa. (That doesn't mean Sun will ship a product using Kawa, though it is not out of the question.) I'll also be working with Chris Oliver and other smart people on language specification.
Previous Topic
