David R. MacIver

More soberly: In a week or two, Go. Then some other languages. And of course a whole bunch of engineering and UX work as people use this thing and inevitably find that it is perfect and wonderful and has no flaws and we just have to spend all our time fielding praise for how great we are. (And maybe fix a few bugs)

Lots of plans for the future of these libraries. This is very much just the start. But I'm pretty pleased with where we are as a start.

@tintvrtkovic@mastodon.social What do you want from me! I wrote https://github.com/DRMacIver/shrinkray It's a totally different testing tool that results from me obsessing about test-case reduction! I have breadth and versatility!

@griotspeak@soc.mod-12.com Today: It'll be a little bit rough, but you could try just porting the rust library. It's not a huge amount of code. Bet Claude or another coding agent could do most of the work. We're going to get the protocol properly documented and to put a bunch more work into making it easy for people to spin up new libraries, but aren't quite there yet.

@csepp @rixx (It's not clear to me that this is what happens though. Also a lot of the value-add can be done less well with open weight models today, and I'd bet on them catching up enough to be good at the drudge work that is where the biggest value add is)

Based on what I've seen from them I would bet that they're mostly pretty close to good implementations (except for the Agda one. I don't think the Agda one works at all. I gave the task to Claude because I like making robots suffer and wanted to see what would happen) but I'd expect they contain at least one absolute howler.

Typescript is next on the official schedule but my guess is we'll probably release a sneaky C++ or OCaml version first.

@griotspeak@soc.mod-12.com Yeah, hand translating also perfectly valid! Give it a go. It won't be entirely straightforward, but it should be a lot easier than a full Hypothesis port.