Re: TypeFamilies vs. FunctionalDependencies & type-level recursion

Hi, 2011/6/21 Simon Peyton-Jones

I suppose that could be changed, yes, but what exactly are we trying to solve here? One can already specify different behavior for constructors with/without named fields. Are we trying to avoid OverlappingInstances? Then yes, this might help, but I'm not sure this change alone would make all generic programming possible without OverlappingInstances.****

** **

To be clear, I wasn’t advocating a change, just saying that there’s no GHC-HQ imperative to avoid them.****

Ah, right. I would also be happy to help improving the current generics mechanism. Also, reading up on OverlappingInstances in the User's Guidehttp://www.haskell.org/ghc/docs/latest/html/users_guide/type-class-extension...(namely the incoherence example without IncoherentInstances) made me see the evilness of the current implementation. But I would only favor deprecating OverlappingInstances when there is a clear alternative supporting (and potentially improving) the current uses for OverlappingInstances. Pedro

Simon****

** **

*From:* José Pedro Magalhães [mailto:jpm@cs.uu.nl] *Sent:* 21 June 2011 09:01 *To:* Simon Peyton-Jones *Cc:* David Mazieres expires 2011-07-21 PDT; oleg@okmij.org; ccshan@rutgers.edu; Dimitrios Vytiniotis; haskell-prime@haskell.org

*Subject:* Re: TypeFamilies vs. FunctionalDependencies & type-level recursion****

** **

Hi,****

2011/6/21 Simon Peyton-Jones ****

| One thing you could do to help in this specific case would be to use a | different M1 tag--e.g., M1 S ... for selectors and M1 NS ... for | fields without selectors (or K1 NS). I presume you've already | considered this and/or it's too late to make such a change. (Or to | move the distinction up to the constructor with two different | constructor tags, CR and CN for record and no-record.)****

I don't think it's too late to make a change. The stuff has only just gone in, so it's still very malleable. There may be other considerations, but legacy code isn't one of them!****

I suppose that could be changed, yes, but what exactly are we trying to solve here? One can already specify different behavior for constructors with/without named fields. Are we trying to avoid OverlappingInstances? Then yes, this might help, but I'm not sure this change alone would make all generic programming possible without OverlappingInstances.

(Also, I always thought UndecidableInstances were "more evil", in some sense, and this change does nothing to remove the use of UndecidableInstances for generic programming.)

Cheers, Pedro****

** **