galonk

Would it be useful to (say, using ML) encrypt text such that the ciphertext looked like intelligible text itself? So you could hide an encrypted message in plain sight? Has this been done?

I maintain a (formerly) popular Python third-party library. At one time it was in the top few hundred most-downloaded packages on PyPI. Unfortunately (thanks to work and life) I haven't actively worked it for several years now, even though it's still a dependency for a few very well-known projects.

Every time I try to get back to it and get started again, I remember all the roadblocks. These are probably not that hard to overcome, but psychologically they keep making it easier to do nothing rather than dig in:

- All the latest work is in an "next generation" branch that's been converted to Python3. I'm not sure whether I should release this as a separate "NG" package, or do a bunch of work to make it 2.x compatible again, or if I can use something like 3to2.

- There's a lot of changes on that "next generation" branch that I need to document.

- I haven't read any of my project email or mailing list traffic in years.

- The project is hosted in mercurial on Bitbucket, I want to migrate it to git on Github. I have an old attempt to transfer it on Github already, I need to figure out how to update that to the latest mercurial revision.

- I should release a new version of the "old" (released) codebase with bugfixes, but I'm not as familiar with it anymore, compared to the "next gen" code.

Has anyone else gone through this? Any advice on where to start? Should I try to find a partner to help motivate me? Even just words of encouragement might help at this point ;)