Never tried Bombadil, but I have used Hegel for property-based testing and can highly recommend it, and it has an excellent Nix story (at least last time I used it, hopefully it’s still good!). I have used it for nix-storage-plugin and it has been very good, and helped my coding agent find several edge cases in its algorithm implementations (especially parsers!) and fix them.
I should use it for more, but most of the new code I (& my agent) write is Nix or personal config stuff, which is harder to test/not really worth it to test. (Maybe if/when I actually start building my ATProto+Iroh git idea…)
The most interesting bit here to reflect on, for me, was the frank discussion of “okay, so, our niche is now super relevant and about to see a massive influx of new people–what do we do?”.
That question is incredibly relevant to our own community, which I think is only a year or two (if that!) from a similar state of affairs.
Wow, if in that first part of the talk could you replace software reliability with reproducible builds (or whatever you care about Nix/Nixpkgs/NixOS supporting), it does resonate pretty well.
Sorry for the tangent, I just really like PBT!
Sometime I should watch the talk in more detail instead of skimming the video.