[bitc-dev] monads

[David-Sarah Hopwood]

Jonathan S. Shapiro wrote:
> On Tue, 2008-03-25 at 11:07 -0400, Sandro Magi wrote:
>> Jonathan S. Shapiro wrote:
>>  > [...]
>>> This leads me to the question: should we add the monad concept to BitC?
>> As an alternative, there is now even a Haskell dialect with an effect 
>> system in place of monads:
>>
>> http://www.haskell.org/haskellwiki/DDC
> 
> Yes, but that doesn't address the objectives that adding a monad system
> would resolve:
> 
>   1. Direct translation of existing haskell code.
>   2. Formal analyzability.
> 
> However, I neglected to consider the impact of lazy vs. eager
> processing, and that is quite a major difference.

A substantial fraction of existing Haskell code does not rely on laziness.
The semantics of Haskell don't actually require an implementation to be
lazy; they allow it to be (although some existing code will of course
run out of memory if evaluated strictly).

-- 
David-Sarah Hopwood