GHC *is* resource hungry

[Ralf Hinze]

> As far as I know, Hugs doesn't hash-cons types, yet it manages to
> typecheck these pathalogical examples in reasonable time/space.  I
> vaguely recall there being a specific modification to Hugs's typechecker
> to handle this, but I can't remember what it was.

Ahem, what do you mean by `pathalogical examples'? This is my
brandnew DSL for constructing trees ... 

Seriously, I think Hugs does not run into this problem simply
because it does not use System Fomega for intermediate representation.
It's the type arguments that kill GHC. On the other hand, the program
is not that big to justify >1.5GB space.

Cheers, Ralf