This definitely feels like it might change how we want to approach an error message catalog. That's purely speculative, but I'm looking forward to learning more (apparently they're releasing a guide to clojure.spec
tomorrow).
At least Rich admitted in the blog post that these errors are a problem for users.
😉
I haven’t had time to look at the announcement in detail, but I’m unconvinced that treating code as data for validation is a good approach.
I mean, Racket has contracts, but they have a different system for macro parsing.
I guess I’m taking a wait-and-see approach. I’ve never looked at Racket contracts (which sound like they were definitely an inspiration here), so I have basically no intuitions on this.