Branching vs. forking

[jeremymanning]

Given that we are now working with a "live" project, I think it's time to stop using branches to store in-progress code. To this end, let's:

• [ ] close existing branches as it becomes possible (we can phase them out gradually to the extent that we won't lose work)
• [ ] moving forward, the preferred method of adding/modifying hypertools is:
  • Fork repo
  • Do the modifications
  • Submit a pull request

Branches should, I think, be reserved for fundamentally different (but self-contained and functional) versions of the code. We could eventually move to something like a stable/experimental/testing branch setup, whereby:

• new features are added to the testing branch
• as they develop, we gradually migrate features from testing (largely non-functional, breaks frequently, incorporates latest features) --> experimental (mostly functional but changes frequently, incorporates all "working" features) --> stable (everything works to the best of our knowledge, stable API with backwards compatibility; this version is what we push to pip)

[jeremymanning]

All non-master branches are now closed.