| | 1 | === Comparison between plan MS and MC === |
| | 2 | |
| | 3 | ==== Plan MC (type functions in GHC) ==== |
| | 4 | |
| | 5 | * local assumptions are turned into rewrite rules |
| | 6 | |
| | 7 | * "simple", easily fits into GHC's current scheme |
| | 8 | |
| | 9 | * but has some restrictions, eg. |
| | 10 | |
| | 11 | - (local) equations must be oriented |
| | 12 | |
| | 13 | - (local) equaitons must be terminating and confluent |
| | 14 | |
| | 15 | - only during constraint generation we may be able to test whether conditions are satisfied |
| | 16 | |
| | 17 | ==== Plan MS ==== |
| | 18 | |
| | 19 | * maps the problem to CHRs |
| | 20 | |
| | 21 | * more complete, less restrictive but may require more substantial changes to GHC's inference engine |
