It is really sad and bad to don't finish something you've began doing by publishing the results so they can be widely available/read/cited/criticized(good ones)

Here I want to publish some results, sketches and may be some code which I hope somebody can find usefull or even interesting to read.

Papers has some comments and correspondence which wouldn't be present in the final draft but for now they can reveal more or what we were thinking/debating about.

JAD - joint approximate diagonalization