Pondering the possibility of using Prolog for bootstrapping a type theory + programming language implementation…
…though I'll probably reconsider once I start hitting all the type and mode errors. 😅
Conversation
I don't see the point. No native Prolog feature can be reused sensibly, we just get a slower, untyped and more verbose transcription of Haskell/ML code.
2
Maybe it is a silly idea… I guess the Prolog feels a bit closer what I would write in a language reference, but is something I can also run. I'd also like to avoid tying a bootstrap implementation to an language that is already hard to bootstrap itself?
1
1
It's easier to bootstrap an untyped LC with constructor matching and without backtracking, than Prolog.
1
Yeah I guess that is true…
1
1
If only there was a way to do that though that was more of a 'mindless transcription' of some declarative rules. In saying that, I guess my Prolog is pretty imperative – and if I went super declarative I probably wouldn't be able to run it! Gah.