Re: [nebula-dev] Future of Nebula

[Edwin Park <esp@xxxxxxxxxx>]

Hi Wim,

So how could we implement what you suggest but still be agile, management-wise.

Here are a few suggestions:

1. Create separate git repos and hudson builds for each widget. This isn't a big deal to set up, and it will give freedom to widget maintainers to focus on their projects without worrying about potentially impacting unrelated widgets. It will also for easier tagging and branching of individual widgets.

2. Revamp the widget catalog. This is the first thing prospective users of Nebula will see, and it should convey accurate and useful information to them.

    a. Indicate who the lead maintainer is for each widget and provide contact info, or if there is no active maintainer indicate this as well. It's great that we have identified this already, but we should put this info into the catalog because it tells users how they can get support. Users are not going to dig through our wiki to find this info.

    b. Provide links to the widget's source repo and ci build (see #1)

    c. Provide links to the bugzilla for each widget so users know how to submit bug reports and maintainers can see what needs to be done.

3. Create separate widget subprojects. This is probably the biggest pain to deal with administratively, but as far as agility is concerned, I think it will make us *more* agile because it will allow widgets to be released independently.

Now I do realize that there are a large number of widgets in Nebula that see very little active development, so creating separate projects for each of them may seem like a waste of time. But what it will do is allow us to release them separately from the widgets that *are* being actively developed. And anyway project setup is a one-time cost. If a project is set up and nothing ever happens with it, fine. There's no additional work that needs to be done. But we never really know when development is going to start happening for a given widget. This way when it does, we'll be ready.

I think we can easily do at least 1 and 2 in the short term, and I'd be happy to help doing them. 3 is more work to set up and for individual work maintainers going forward, but we can have all of the projects operate similarly and use shared templates and things to minimize the effort involved here.

Edwin