You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by Francois Letellier <fr...@objectweb.org> on 2005/08/26 12:17:22 UTC

Re: [proposal] Oscar OSGi Project



>Enrique Rodriguez wrote:
>
> > Sylvain Wallez wrote:
> >
> >> The problem here is that ObjectWeb is a non-profit organization that
> >> hosts a large number of project, a bit like the ASF. It's not like an
> >> independent group (log4j, spamassasin, ibatis, etc) with a single
> >> project that moves to the ASF for the community and legal oversight
> >> it provides.
> >
> >
> > You can also say that "SourceForge is a non-profit organization that
> > hosts a large number of projects" and that "Oscar is an independent
> > group with a single project that moves to the ASF for the community and
> > legal oversight it provides."  In the close to 1 year that I've been
> > working with Oscar, I never got the impression that the Oscar-ObjectWeb
> > relationship was anything more than infrastructure.
>
>
>Hmm... comparing ObjectWeb and SourceForge isn't really fair!
>
>Note that I'm in no way affiliated to ObjectWeb, but I happen to know a
>few people there and they raised some issues when I mentioned the new
>Oscar ASF project on my blog (see comments in [1]) and during private
>discussions.
>
>Now if they really have a problem with what's happening, it's up to them
>to react. So let's move on!

All,

I'm quite surprised by this thread, that I'm just discovering today. I 
don't monitor Apache's mailing lists all the time, and this explains the delay.

We (ObjectWeb) brought up the subject of OSCAR "moving" to Apache weeks 
ago, if not months.
Due to good relations between our communities, and especially between 
individuals in these communities, some people in ObjectWeb prefered to use 
private discussions to express suprise and concern and to try to better 
understand what was going on.

In the ObjectWeb approval process for new projects, attention is paid to 
whether similar projects exist in other organizations. We think it's basic 
courtesy, in addition to pragmatic thinking. Open source is about sharing 
and reusing. Maybe Apache Incubator's approval process may follow similar 
rules.

OSCAR has been hosted at ObjectWeb for some time now. Rick, the project 
leader, also has been hosted by consortium members. He sits on ObjectWeb's 
College of Architects. I remember listening to him at some conferences 
organized by ObjectWeb members, including the "DECOR" conference, where he 
was invited as a keynote speaker. OSGi also is considered an option for 
refactoring some ObjectWeb projects, including JOnAS. Definitely the 
OSCAR-OW relationship has extended beyond "infrastructure".

Nothing prevents contributors to Apache projects to also contribute to 
ObjectWeb projects. We grant access to any committer a project leader 
choses, regardless of their affiliation. In addition, OSCAR is under a BSD 
license. This means that Apache projects can take dependencies on OSCAR 
without compromizing ASF's license policy.

Mike Milinkovich (Eclipse) wrote on this mailing list 
(http://incubator.apache.org/mail/oscar-dev/200508):
"Apache and Eclipse (and ObjectWeb for that matter) have a pretty good 
track record at co-operating at many different levels. Why it would serve 
the interests of Apache, Eclipse, ObjectWeb, OSGi or the greater open 
source community to have directly competing projects is beyond me." I 
couldn't agree more with Mike.

Now that an OSCAR (Felix?) incubated project has been created at Apache, I 
can't imagine it being promoted as top level before this issue is settled. 
The best way to go would probably be to find a solution where a single 
project would benefit from the support of the two communities (Apache & 
ObjectWeb). Maybe the notion of "sister projects" would apply here: Apache 
OSCAR would mirror the existing ObjectWeb OSCAR.

Here, we have an opportunity to let our two communities work together. I 
think we should seize this opportunity for everybody's benefit.

Best,




>Sylvain
>
>[1] http://www.anyware-tech.com/blogs/sylvain/archives/000204.html
>
>--
>Sylvain Wallez                        Anyware Technologies
>http://people.apache.org/~sylvain     http://www.anyware-tech.com
>Apache Software Foundation Member     Research & Technology Director


François LETELLIER
INRIA Rhône-Alpes - ObjectWeb Consortium Executive Committee
email: francois.letellier at objectweb.org - cell: +33 (0)6 84 64 00 24 - 
Skype: francois.letellier
blog: http://os3g.blogspot.com/ ObjectWeb: http://www.objectweb.org

Re: [proposal] Oscar OSGi Project

Posted by Francois Letellier <fr...@objectweb.org>.
At 13:37 26/08/2005, Niclas Hedhman wrote:
>On Friday 26 August 2005 18:17, Francois Letellier wrote:
> > The best way to go would probably be to find a solution where a single
> > project would benefit from the support of the two communities (Apache &
> > ObjectWeb). Maybe the notion of "sister projects" would apply here: Apache
> > OSCAR would mirror the existing ObjectWeb OSCAR.
>
>I would be interested to hear how you envision this to be setup. To me it
>sounds interesting, yet don't understand how it could been made to work
>practically.

What about a "sister project" (see eg 
http://www.apache.org/foundation/press/pr_1999_06_30.html)?



>Cheers
>Niclas

François LETELLIER
INRIA Rhône-Alpes - ObjectWeb Consortium Executive Committee
email: francois.letellier at objectweb.org - cell: +33 (0)6 84 64 00 24 - 
Skype: francois.letellier
blog: http://os3g.blogspot.com/  ObjectWeb: http://www.objectweb.org


Re: [proposal] Oscar OSGi Project

Posted by Niclas Hedhman <ni...@hedhman.org>.
On Friday 26 August 2005 18:17, Francois Letellier wrote:
> The best way to go would probably be to find a solution where a single
> project would benefit from the support of the two communities (Apache &
> ObjectWeb). Maybe the notion of "sister projects" would apply here: Apache
> OSCAR would mirror the existing ObjectWeb OSCAR.

I would be interested to hear how you envision this to be setup. To me it 
sounds interesting, yet don't understand how it could been made to work 
practically.


Cheers
Niclas