[bitc-dev] constraining bitc's type system

[Swaroop Sridhar]

Sam Mason wrote:
> Maybe I just like the low level control and the extra memory to store
> all these versions and a bit more cache pressure to get them back
> seems import.  How much this matters in reality I don't know though.
> 
> I'm only suggesting that the specific members that you want to use
> polymorphically need to be boxed and if you're not writing completely
> generic code not even that restriction need apply.   It's at this point
> I realise that it wouldn't be polymorphism any more.  Humm.  I still
> think it would be nice to be able to know that specific patterns wouldn't
> result in polyinstantiation and so would have more predictable runtime
> behaviour and the bonus of being able to interface with external code
> without too much work.

Actually, I think it is not difficult to get what you want even in the 
existing system. Prof. Shapiro had suggested a while ago that we should 
have a special typeclass (say refType) to which all reference types 
belong (that is, the compiler will automatically add relevant 
instances). Now, you can write:

(define (f x:(forall ((refType 'a)) 'a) x)

This will ensure that the polymorphism is constrained to ref-types only, 
and you can be sure that there will be no polyinstantiation due to the 
polymorphism of 'a.

Swaroop.