Hi Scott,
Great to hear that the January is in use out there. Thanks for following the list on this incubating project to keep up with the changes.
We currently provide the plug-ins as top level entities, like other foundational stuff (e.g. Orbit) without the features. Perhaps this is wrong because it's not what consumers expect. But as January is not consumed directly by end users I am not sure what a feature with a single plug-in adds.
My understanding is that the biggest problem we are facing is twofold (but related)
1 - contributors to org.eclipse.january plug-in, which has very few dependencies, are put off by importing the January git repo and finding out they need things like xtend compiler installed in their dev environment and that their are a whole bunch of unexpected dependencies (like UI and Xtext ones).
2 - consumers have the same problem, if they don't know to add just org.eclipse.january plug-in to their target, they end up with loads of unexpected dependencies too.
What I would love is a concrete solution for these. My current inclination is that we should follow Xtext's lead and split up the git repo and p2 site. All under the January name and leadership, but ease the problems identified in this thread. Something along this line seems to give the best of both worlds. The other possibility is just better docs + perhaps two Oomph setups, one core and one everything.
I truly believe we can make this a one stop shop for the Science data structures. Now we just have to find tine to do one of these solutions.
Jonah