Yes agreed but I think if we handle some of the issues on that confluence page
our vision will be clearer than it is now and we can properly evaluate this technology.
At that point we'll see just how much overhead it will incur.
Although I think that OSGi is a cool stuff and we should adopt to it
at some point, I don't think it's the right time at this point. Moving
to OSGi is a big architectural change and it will effect many parts of
the server. So we should all be involved in it. However we have more
urgent matters now on our plate like Replication. I think the OSGi
effort should continue in a sandbox and make sure that it will fit
into the server smoothly. So I think we should wait until we clear our
On 5/7/07, Emmanuel Lecharny <firstname.lastname@example.org> wrote:
> Chris Custine a écrit :
> > I think this depends on the roadmap features and whether OSGi will be of
> > benefit in implementing them in the short term. In the long term, I
> > think
> > modularizing with OSGi will be a great benefit to ApacheDS. In the short
> > term, I think it might be somewhat of a distraction and require too many
> > major changes all at one time. I think it could be done this way, but I
> > think my personal opinion of it would be to implement a new round of
> > features first and then work on OSGi after the first round...
> Hi Chris,
> I hope your flight back to USA was just perfect !
> Enrique has set an almost perfect page about the OSGi effort, I think we
> should dedicate as much possible time on this item if we decide to go
> for it. John Colon and Enrique Rodriguez have started working on all
> that, and we have slowed down their efforts (sorry about that John)
> because of the 1.5.0 release we were planning.
> But now, I think as soon as all the issues listed by Enrique
> are solved, we should be ready to move on.
> Enrique, what would be cool would be to add a page where you explain in
> detail how to move to OSGi, with a sample based on a simple module (with
> code and comments).
> Wdyt ?