My analysis for this is the following:
> Widget authors are pretty much individualists. We
have many widgets and many authors but there is not sense
of a "one product" like you would have for other projects.
Every widget is more or less self sustained. Hence there
is no "us".
> Most widgets are feature complete. They have a
described function and there is no reason to evolve this
from year to year.
I agree with that. Nebula is just an "aggregation" of individual
widgets. It's not a real project (refering to the usual definition
of project) since there is no global roadmap and objectives. 1
widget == 1 (mini) project.
But the current approach is actually working well since it fits
contributors and consumers wishes.
> Feedback from the external community (consumers).
I hear of many people consuming Nebula widgets, but there
is not a desire or need to provide feedback apart from the
occasional bug request.
I agree with that. Nebula widgets are usually good quality and we
can find documentations and snippet about how to use them. I guess a
majority of the Eclipse-based RCP apps have used Nebula widget. I'm
also sure that if a "complete" widget has a bug and needs a
contribution, we can easily find someone to fix it (inside or
outside of the known contributor set).
Conclusion: Having these thoughts, I think we
are better of being as we were, a perpetual incubation
project.
Incubation is actually a cool status, it makes things easier
sometimes. For Nebula, I also don't think there is a big benefit in
turning it into a released project. But there is no issue in being a
release neither...
I was also thinking to broaden the scope of Nebula and make
it a home for widgets of other technologies like Swing,
JFace, RAP and _javascript_.
That's a good idea.
I hope in most cases, SWT/JFace based widget would work with RAP. By
the way, we should try to provide an RAP version of the Nebula
examples.
About Swing and _javascript_ in general, I'm not sure Nebula (and
Eclipse in general) is the place where contributors would like to
contribute the most...
However, it would make a lot of sense to make Nebula host some
high-level widgets dedicated to Orion.
I would vote 0.
What would you improve in Nebula?
Showcase it on RAP.
Do you think the project is still relevant?
Yes, definitely.
What do you think about adding other
technologies besides SWT?
I would vote +1 for Orion-based stuff, -1 for the other ones.
Also, Wim, I hope you'll submit an overview of Nebula for EclipseCon
France (deadline is April 15th) ;)
|