[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [nebula-dev] Restructuring Nebula Project
|
Hi everyone,
There is already an example plugin in CVS which creates a view with all the existing Nebula widgets. A few ones are missing (Emil's ones I think). We could easily turn this plugin into a stand-alone application. The website should also be updated to let everyone know that this plugin exists.
I totaly agree with spliting the nebula project into subprojects, but we should ensure we'll still have the possibility to work on each other's widgets (ability to work as a team).
For the release check-list (we need that) and other thing we can use the Eclipse wiki (
http://wiki.eclipse.org/Nebula). I think it would be better to have a summary of what we said on this mailing list (review results, plans, ideas) because this would give a better visibility for the whole project (no need to digg in the list :) ). I've already added some pages for the gallery widget and a link to the automatic IP log.
Thanks Tom for the Gallery build. Can you tell me how does it works and where the resulting file is located ?
--
Nicolas
On Fri, Nov 21, 2008 at 1:07 PM, Tom Schindl
<listom@xxxxxxxxxxxxxxx> wrote:
Hi Emil,
Emil Crumhorn schrieb:
> I'll do a point-by-point reply.
>
> a). Do we really have an option even if it's a lot of work unless we
> convert to sub-project? Combining everyone's widgets and coordinating
> would never work I think, and as more projects join nebula it would be
> even harder. If we don't convert, we'll stay where we are now and go
> nowhere, or we do a lot of work and go somewhere. I vote for the latter.
>
> b). I guess this depends on a)
>
> c). I don't see why not. I don't like that it's potentially "a pain in
> the butt" to join Nebula though. I can see why some people go the
> sourceforge route because there it's easy to control your project and
> you can manage it quite easily as well. Setting it up is also mostly a
> no brainer. I don't think people are against a bit of work, but if it
> becomes too much, they're going to wonder (like I did), if it's really
> worth the effort. Also, sourceforge has more of a "these people are
> working on the project", but I never see that in Nebula in the same way,
> and people can't easily join to contribute. I know there's a reason
> there are barriers, but sometimes they're too hight.... But I'll still
> say "yes".
>
> d). Agreed on the API. But I think some widgets wouldn't mind a 1.0
> stamp as their api hasn't changed in a while and they're quite stable. I
> don't know if there should be a review process first though to ensure
> they conform to API and Widget "standards", probably would be good, even
> if it's just a "hey guys, could you take a look at ......".
>
You are completely right and to get your widget released you need to go
through a release review! I plan to set up a check list what I expect
from a widget to get released (some of the following things come to my
mind):
- Naturally an automated signed build
- A clean and concise API
- Warning free
- Meaning full OSGi-Information (Versioning, Package Exports, ...)
- I18N-capable (e.g. no untranslated Strings for errors, ...)
- Fully JavaDoc API (public/protected methods and fields)
- A set of 2-3 Snippets showing how to use the widget
We would even more harm the Nebula project if we release widgets not
ready for it and I'd never +1 a release before I'm confident with the
widgets state.
Do you guys think we as a nebula project should start thinking about a
small "Example-Project" contribution we can make which creates a small
RCP-Application showing the cool controls we have?
Tom
--
Nicolas Richeton
