[Haskell-cafe] Re: lawless instances of Functor

[Heinrich Apfelmus]

Dan Piponi wrote:
> Derek Elkins wrote:
> 
>> Yes, I have the same problem...Basically, I'm
>> pretty sure the construction of that free theorem doesn't rely on any
>> of the actual details...
> 
> For a long time I've thought such a higher order free theorem must
> exist, and I've mentioned it to a few people, and searched hard for a
> paper on it, but I haven't seen an actual statement and proof.
> 
>> At this point, though, I haven't put
>> much effort into proving that the free theorem holds uniformly
> 
> Well I encourage you to as I've a hunch the correctly generalised
> theorem will be quite pretty. I'd have a go but the style of proof for
> these sorts of things is outside of my domain of
> confidence/experience.

This looks relevant:

  Janis Voigtländer.
  Free Theorems Involving Type Constructor Classes.
  http://wwwtcs.inf.tu-dresden.de/~voigt/icfp09.pdf



Regards,
Heinrich Apfelmus

--
http://apfelmus.nfshost.com