Post

Increasingly convinced that specs are now the source and that code is an artifact akin to a lockfile.

*Crashes through skylight* did someone say formal methods?!

wait wait no I take it back everybody

Only half serious, but a lot of the practice of formal specification boils down to "how much can we simplify our model of the system and still have something useful enough for analysis?" Like do we have to specify that a particular batch job runs nightly, or is "it runs sometimes" enough?

Another hot topic is studying whether a codebase matches a specification or not, or if one spec is a more detailed version of another. I've not done a lot of this but I know @dominiktornow.bsky.social has a three step process: formal spec to more-detailed formal spec to implementation code

I should disclaimer that the way we FM folk think about "specifications" doesn't naturally fit into how people intuitively think of "LLM specifications", so I wouldn't be surprised if the main outcome is "inspiration for similar approaches", with the actual techniques developing independently