Re: [Haskell] Proposal for a Standard of Abstract Collections (with Reference Implementation)

It looks interesting and I'm still looking at it, although I think many of the language extensions need to be better thought out. But it exhibits the "creeping Eq" problem: your hierarchy starts class (Eq (coll a), Eq a) => Collection coll a where ... If this is to replace lists, this is unacceptable: I can't have lists of (say) functions? Collection classes should not require Eq instances on the members, except when necessary! Peace, Dylan