The difference with terms is that the arity of type families matters. F1 has arity 1, while F2 has arity 0. The real question is: where is the k bound. In F1, k is bound "before the colon", while in F2, it's bound after
#11719: Cannot use higher-rank kinds with type families -------------------------------------+------------------------------------- Reporter: ocharles | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.8.1 Component: Compiler (Type | Version: 8.0.1-rc2 checker) | Resolution: | Keywords: TypeInType Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: GHC rejects | Test Case: valid program | dependent/should_compile/T11719 Blocked By: | Blocking: Related Tickets: #13913 | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by simonpj): * cc: csongor.kiss14@…, s.eisenbach@…, t.field@… (added) Comment: the colon. Tricky stuff! Could we document this in the user manual? If we end up adopting Csongor's ideas about unsaturated type families (when the matchability shows up in the kind), we'll need to adopt some notation for the forall-case too. I can't say I like `forall k '. k -> Type`! But we clearly need ''something''.
Perversely, type families can pattern-match only on unmatchable arguments. Perhaps a name change is in order.
That is indeed perverse. Worth considering as part of the same effort. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11719#comment:20 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler