H-core (was: Re: ~ patterns)

Malcolm Wallace Malcolm.Wallace at cs.york.ac.uk
Thu Feb 2 06:19:09 EST 2006


John Meacham <john at repetae.net> writes:

> On Wed, Feb 01, 2006 at 02:51:08PM +0000, Philippa Cowderoy wrote:
> > specifying a dictionary-passing implementation of typeclasses - 
> 
> I am thinking we don't specify any particular translation scheme. just a
> sudset of the language that is considered 'core' that every haskell
> program could _potentially_ be reduced to.

There is a potential confusion here between the H-core language as used
in the Haskell'98 report, and ghc's current external-core language.
The former contains classes, and the latter does not.  Ghc-core has
type-lambdas, but H-core does not.  Ghc-core has unboxed values,
H-core does not.

The main conceptual difference is that H-core aims merely for simpler
expression of common syntactic constructs, whereas ghc-core aims
for low-level optimisability.  H-core should be programmer-usable,
whilst ghc-core is intended for automated tools to use.

Regards,
    Malcolm


More information about the Haskell-prime mailing list