Bringing License arguments to Sun

Leo Simons mail at
Wed Aug 23 11:22:42 UTC 2006

(I looked at for a
feedback button or something but can't find it. Thanks for listening

On Sat, Aug 19, 2006 at 07:38:36PM -0700, Stefano Mazzocchi wrote:
[what license should Sun use to open source java]
> I'll bite: the MIT license.

+1, for all the reasons Stefano described. Along with the neccessary,
explicit, relevant patent grants, preferably with GPL-compatible terms
(eg non-reciprocical; would probably automatically meet requirements
off standards bodies and open source orgs worldwide).

The most important concern as I see it in terms of open source java
license choice is to de-fragment the ecosystem as much as possible.
If sun makes haste, I can still see java becoming the #1 language for
desktop applications 'n stuff, and I can still see us all surviving
the longhorn assault that's coming. Dalibor's written a lot of good
stuff on his blog:

(I would paint a little different history lesson and I don't like
dual licensing as a long-term strategy, but other than that its spot

Now, to dwell on alternatives. My personal opinions.

If MIT license is not possible, next up is ASLv2 since it will
still mean (after GPLv3 finalizes) that code can be incorporated
downstream in both apache (harmony and other projects, think
jakarta commons, xml commons, yoko) and GNU (classpath and other
projects like gcj).

If ASLv2 is not possible the next preference up is probably a
custom license that still allows pieces to be mixed and matched
with both the "Free Software" and "Open Source" parts of the
ecosystem. I imagine any such license would still come somewhat
close to ASLv2, and I'll highlight how painful introducing new
licenses is for the entire ecosystem. Please don't do it again :)

If that is not possible, I'll go and be an "ASF heretic" and say
it should be GPLv3. That way, at least healthy co-operation
between the GNU communities and "jonathan java" will still be
possible. In a way it'd be tough luck for harmony, since we've pretty
much determined already we wouldn't be able to depend on GPLv3
pieces, but it would still allow a fruitful "downstream merge"
with "jonathan java" and classpath incorporating the best harmony
bits. The bigger ecosystem still wins. And most "apache java" people
will have no problem making use of GPLed stuff - we're all GCC
users already for example ;).

If that is not possible, I guess the last realistic alternative is
the CDDL. I like the CDDL and I use it for some of my hobby
projects. Its "category B" in the ASF 3rd party license policy
draft [1], so it would still allow some kind of co-operation between
"jonathan java" and harmony. **But** IIUC its still undecided whether
the GPLv3 will be compatible with the CDDL, which potentially means
closing off -- rather permanently -- a really big part of the FLOSS
java community. In this picture, I see all java being ripped out of
OpenOffice (or OO.o suffering from forks and communit dissent),
Java-GNOME and the like not being used, in favor of mono, and java
permanently losing a position on the desktop of anyone but java
developers themselves. Sure, it'll still thrive on the server and
open source java will still be loads of fun, but I'd be removing the
"swing" keyword off my resume.

Last legal point -- I'm guessing all license choices are somehow ok to
the opensolaris community -- GPL is by virtue of having GCC and gnome
and stuff around, CDDL is obviously, and ASL is too by virtue of having
the web server on there.

You know, I'm just not too fussed. You guys will get it right "enough". seems to be "working as intended" these days,
is shaping up nicely, even OO.o is seeing more and more of an actual
community. Listen to all the stuff Geir has to say about the JCP and
opening that up further and then we'll be fine. The jini stuff coming to
apache is a further nice exercise.

Its not just about those 400-or-so open source developers who will be
working on java. Its about all the weird stuff you're not planning or
expecting that will rock the industry; providing rocket fuel to the
crazy rocket scientists. The proper port to BSD might unearth a 10 year
old bug than when fixed means a 15% performance improvement on linux.
Who knows.

The network is the computer, and the community is the real network.

I know I know, preaching to the choir, but somehow this is all being
de-emphasized. Open source is cool, and open source java will be, too!

My main personal interest is that I'll somehow be able to bootstrap a
system all the way from bootstrapping GCC up to big J2EE-based systems
while doing all sorts of interesting measurements (Re: gump) and thus
improve the stability/maintainability of the whole stack. But I'm a
little weird like that. All I need now is open source oracle...


[1] --

More information about the Classpath mailing list