On 02/10/2016 06:30 PM, Joachim Durchholz wrote: [--snip--]
I am not sure how relevant that is going to be in practice. Maybe it's possible to come up with a short "for now" explanation that is consistent with what students will experience; if that's possible, I'd expect that to be much preferrable over a simplified/lying Prelude.
One thing about a simplified Prelude: It will profoundly unnerve newbies. The feeling of shifting ground the first time they see that something is amiss will make them feel insecure, because they won't know how much of what they already learned will have to be thrown overboard.
How do you know that they will feel this way? I feel like most people in this thread are just going by assumptions. (I guess I could be accused of the same thing, but I guess arguing for the status quo doesn't really incur a burden of proof. We can see that it at least works somewhat well. Granted the $ type is new, but advanced type signatures abound in Haskell already. It's kind of par for the course.) Regards,