H-core (was: Re: ~ patterns)

[Robert Dockins]

On Feb 1, 2006, at 5:12 AM, Malcolm Wallace wrote:

> Taral <taralx at gmail.com> writes:
>
>> On 1/31/06, Simon Marlow <simonmar at microsoft.com> wrote:
>>> I've been swayed by the arguments put forward by the ~- 
>>> proponents, so
>>> I'm not going to champion the removal of ~ any more.
>>>
>>> We must find *something* to throw away though! :-)
>>
>> I still like the idea of splitting Haskell' into Haskell'-core and
>> Haskell'-lazy, and moving ~ and ! patterns into Haskell'-lazy.
>
> The Haskell'98 Report already uses an informal notion of a "core"
> language, into which other syntactic constructs are translated.
>
> I think in Haskell-prime we ought to define this core precisely
> and formally.

I'd like to second this.

> One can even imagine someone developing a pure H-core compiler, with
> the fuller language implemented as a pre-processor over the top!
> (I know at least one person who would prefer to write programs in
> core rather than Haskell'98...)

In light of the recent post on optimizing core, it seems like this  
might be a very good way to allow people to optimize their inner  
loops without having to trick their favorite compiler to do the  
optimizations they want.


Additionally, a standard for core would allow a new level of tool  
interoperability.  Haskell front ends and backends could be cleanly  
separated along a well-defined border.  DrIFT and Haddock and others  
as well could benefit.  Happy could generate core directly.... anyway  
you get the idea.



Rob Dockins

Speak softly and drive a Sherman tank.
Laugh hard; it's a long way to the bank.
           -- TMBG