User loginNavigation |
Unifying map and mapM through subtyping?In a Haskell-like language, could we unify It seems that such conversions aren't as problematic as a general subtyping system, since proving confluence isn't a problem. We'd have to do something to prevent the typechecker from looping indefinitely trying to unify Are there other difficulties that this approach might cause? What sort of trickery might be needed to convince an optimizer over a typed intermediate language that the Identity introduction and elimination functions should have no operational effect? (Furthermore, is it true that they should have no effect, or is there a useful distinction between By Douglas McClean at 2010-02-26 02:58 | LtU Forum | previous forum topic | next forum topic | other blogs | 4860 reads
|
Browse archives
Active forum topics
|
Recent comments
9 weeks 5 days ago
13 weeks 6 days ago
15 weeks 4 days ago
15 weeks 4 days ago
18 weeks 2 days ago
22 weeks 6 days ago
22 weeks 6 days ago
23 weeks 2 days ago
23 weeks 2 days ago
26 weeks 1 day ago