I find the stuff on jneem’s github wayyy more intelligible than the actual current pijul code. Granted I think it should be largely equivalent!

Full disclosure : totally planning to messing around with a Haskell port of some of these ideas sometime soon.

I read that merging blog post a while back but didn't completely grok it, I've yet to read the paper though. Do you know of any other write-ups on patch commutation?

Unfortunately, no. It's pretty much my only exposure to the idea.

A shame, thanks for the reply though. I'll set aside some time this weekend and read through it again.