Point: "Write once run anywhere" means there can only be one J2EE JVM... to do otherwise undercuts Java vs .Net. The existing open source JVMs are all slightly (or not so slightly) different in the wild... ISV Hell. So I propose this discussion should begin with "given that there can be only one Java..."
Counterpoint: But there are already several JVMs. IBM has their own JVM's tweaked to work on their hardware, for example... it doesn't make sense to expect Sun to optimize Java for IBM platforms like POWER, z/OS or AS400, does it? And there are other JVMs as well, but ISV's still develop against Java. And besides, even if there was just one open source JVM, who would pay the big bucks to run it through the compatibility test suite? Wouldn't they have too much influence over the project, since they were paying for the ride?
Point: A consortium like OSDL with hundreds of members could do that. But what about Sun? They arguably control the JCP, however subtly. Does it matter what a consortium developing the code and paying for the CTS does? I mean, the open source group couldn't just go their own way, scratching their own itches. They couldn't even call their VM "Java"... Sun owns the Java trademark. Where does that leave them?
Counterpoint: The JCP/JSR system should be replaced by by a "Benevolent Dictator" like Linus.
Point: Who do you have in mind? Besides, JSR99 will allow different implementations.
Counterpoint: It's still part of the JCP.
That's where I give up...
-Rayme.
On Apr 9, 2005, at 6:48 PM, [EMAIL PROTECTED] wrote:
Phillip Rhodes wrote:
[EMAIL PROTECTED] wrote:
it may be possible to actually fix these bugs for the broader Java community.
Scwartz is even hinting that Java may one day be open source too! This would be so sweet and make too much sense to be reality.
Not to start another religious war over "Open Source Java" but...
does it even matter if Sun open-sources their JVM or not? The language
of the JCP specs, if I understand correctly, allows compatible open-source implementations; and there are already several projects underway to develop open-source JVM's and Java class libraries.
Yes it does.
See Kaffe, GCJ, SableVM, JikesRVM, ElectricalFire, and probably others (and yes, I'm aware that ElectricalFire is not really active right now, but there is code out there that somebody could pick up and run with).Kaffe....sucks. GCJ is not a VM and Java as a langauge without the runtime environment is not a very compelling story. And none of the others actually are much of an effort.
I suggest that anyone who is enamored with the idea of open-source JavaThey are irrelevant. So long as the TCK and the reference version are closed source the only thing you can do is create POOR copies. It took 20 years for Linux to emerge.
should contribute to one of those projects, or GNU Classpath, and quit
worrying about what Sun does. If they open their code eventually, great. If not, the F/OSS community will carry on without them.
Eventually I'll have my own VM project underway (however I'm deeply entrenched in JBoss Mail Server and a new POI project ATM) with very different goals than Java
(more similar to the .NET VM but a bit more ambitious perhaps). Personally, if they don't open up Java and participation in Java then I say we learn from Java and move on.
So long as the platform is closed and locked in these largely theoretical specification committees, it is irrelevant that you can create implementations of it.
Finally, a proper JIT'd VM is simply hard work. Its not feasible to do that as a "catch up" (in a way that you could both apply for and pass the TCK) and then Java is a few years
behind the cutting edge of research in that field to start with.
Sun's present closed system is a recipie for a fork of Java not for an open source edition. Its exactly the opposite of their story.
-Andy
PS this is probably the exact opposite opinion of my employer :-)
TTYL,
Phil
_______________________________________________ Juglist mailing list [email protected] http://trijug.org/mailman/listinfo/juglist_trijug.org
_______________________________________________ Juglist mailing list [email protected] http://trijug.org/mailman/listinfo/juglist_trijug.org
