> It strikes me that using Bits instead of Enum might be more likely to be what > people want in many cases
It wouldn't hurt for you (or someone) to implement this. I'm sure it would be useful. On Wed, Aug 12, 2009 at 2:09 AM, Ketil Malde<[email protected]> wrote: > Felipe Lessa <[email protected]> writes: > >>> There are some funky Enum instances around: > >> IMO it's implicit that keys overwrite eachother whenever their >> 'fromEnum' is equal, however that may be spoken in the docs. > > I couldn't find anything explicit in the documentation. I'd suggest a > clear note at the top, dismissing the (IMO natural) notion that > "EnumMap k v" behaves like "Map k v" (which was true for IntMap and > Map Int, I believe). > > And perhaps also note that you will get exceptions for values outside > the Enum range. > > It strikes me that using Bits instead of Enum might be more likely to > be what people want in many cases - but perhaps that would be too slow? > Also that Enum really should map to Integer, but again, that's a speed > issue.¹ > > One could also question the sanity of using e.g. floating point values > as keys, but Map supports this, so who am I to judge. > > (Also, a minor documentation niggle is that Haskell only guarantees 30 > bits for an Int, it's GHC that uses Ints of 32 and 64 bits. One could > argue that it is the report that should be fixed here, unless one can > imagine a program that depends on correct modulo arithmetic with an > unknown quotient.) > > -k > > ¹ This of course migth give the careless reader of the Report the > impression that the Haskell community values speed over correctness, > and thus that we actually are aiming for popularity and mainstream > recognition after all. > -- > If I haven't seen further, it is by standing in the footprints of giants > _______________________________________________ > Haskell-Cafe mailing list > [email protected] > http://www.haskell.org/mailman/listinfo/haskell-cafe > _______________________________________________ Haskell-Cafe mailing list [email protected] http://www.haskell.org/mailman/listinfo/haskell-cafe
