On Tue, Apr 04, 2006 at 02:53:36PM -0700, Andy Gill wrote:
Another issue is that being able to detect cyclic structures would make it impossible to express deepSeq as a Haskell -> Haskell translation. which is no good.
I am trying to understand this requirement. For the sake of what must all primitives be expressible as a Haskell -> Haskell translation?
Mainly it is an excellent proof that no undue burden is being placed on any implementation, current or future. It also gives a way to reason about its behavior and is a way to ensure you don't accidentally miss defining any behavior or break referential transparency or any of the other properties haskell programmers expect. not that it has to be implemented via the translation of course. things like DeepSeq and Typeable will most likely have optimized versions on various compilers which is why I'd like to see the restriction that the only way to create instances for these two classes is via the "deriving" mechanism of the compiler. for the record, jhc can do a super optimized Typeable, but not a DeepSeq, so will likely have to use the standard class definition of DeepSeq (which it can already derive, under a different name). John -- John Meacham - ⑆repetae.net⑆john⑈