28: Richard Eisenberg
The Haskell Interlude
00:00
Haskell's Type Inference: A Non-Canon Call
No active plans right now to get rid of that, although we shouldn't have it. The reason I say it shouldn't really exist is that there are constraints that are sort of fresh that have just been created. And then there's the constraints that we've already processed. But actually, in almost all cases, we know whether something has been canonicalized or not. So non-canonical should really just be out in its own type.
Transcript
Play full episode
