"it's probably a good idea to first decide what the actual needs are"
Some ideas on that...
- SDL2 provides better platform support in some places, and some nicer features.
- SDL1 provides better platform support in some places.
- pygame apps would like to continue working.
Things like mobile support, I agree are important, but I guess getting the basics done first is more important. Everything else can flow from there.
Currently we don't even have a good story for distribution on pypi,
windows, linux, or mac (The platforms where people make things on). I
think we should keep our goals humble, and achievable. Complete them, and move on to the next steps.
Daniel Pope had a good point about not putting in any experimental features, and to keep it simple. We've learned that lesson, and have removed things already that were not maintainable. It's the approach that CPython now takes as well - things need to gain popularity and prove stability before they are moved in. Additionally now it's much easier to use external packages via pip, so bundling all together is not as useful as it once was (binary distribution is still challenging however).
The app distribution work(so people can distribute their own games easily), I think should continue with the wider python community. A tool I'm working on has had massive contributions from someone who doesn't use pygame, or game tools at all. I think that Thomas Kluver probably has the same experiences with his packaging work. Trying to get things fixed in the python distribution tools themselves has proved challenging, but has also proved to sometimes work.
With this proposal, we should be able to work on other things like mobile, and extended graphics support afterwards. I agree that they are important. There's also nothing stopping people from working on them separately(or raising resources to fund that work).