What's the future of OpenJDK?

[Andrew John Hughes]

On 2 March 2010 14:08, Mark Wielaard <mark at klomp.org> wrote:
> On Tue, 2010-03-02 at 14:26 +0100, Volker Simonis wrote:
>> I can only agree with Mark Wielaard and his call to "Endorse and
>> promote individuals, organizations and
>> companies to contribute and encourage them to work together", although
>> I think this will be a lot harder for the HotSpot VM case than for the
>> J2SE libraries.
>
> That is funny because I think it is the other way around actually.
> Although technically challenging, in the end HotSpot is "just" a VM for
> which there are pretty tightly defined constraints on how it should
> work. We know how to replace it completely. See the Cacao integration in
> IcedTea. And we have people hacking on significant projects like Zero
> and Shark which replace huge chunks of it.
>

There's a big difference between a working VM and a highly performant
VM.  It's one thing to keep the VM building and running with the
current class libraries, and quite another to add new optimisations,
improved garbage collection, etc.  If the GPLed HotSpot was to be
abandoned by Oracle (which is jumping the gun a bit as we don't really
know anything yet), one would hope that their absence may lead to
increased interest from other parties, but it wouldn't be possible to
do anything but basic maintenance on HotSpot with the current tiny
OpenJDK community.

> The core class libraries on the other hand are much more fuzzy, don't
> have a very strict spec, and are mainly defined by current behavior of
> the openjdk implementation. Changing anything here quickly runs into
> either hard to judge "compatibility" discussions or politics over
> how/when to extend the public api. Both of which are pretty much
> impossible to test independently since the testsuite is basically "all
> the programs ever written out there" which depend on all kinds of weird
> quirks of the current implementation (we had to work pretty hard to be
> bug compatible in GNU Classpath).
>

I agree the class libraries are even worse; without even bringing in
GNU Classpath, there have been numerous divergences between OpenJDK
and the proprietary Sun JDK.

> So from my perspective hacking on the VM is much easier since there are
> already so many people out there who have hacked on various parts and
> pieces that show that things will just work. While replacing parts of
> the core class libraries has much higher risks of breaking stuff out
> there.
>
> Cheers,
>
> Mark
>
>



-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8