Re: lazyness

Adrian Hey <ahey@[EMAIL PROTECTED]
> writes:
> That said, I think this approach does lead to some nasty practical
> problems and ambiguity with respect to space and time behaviour of
> real programs (which is a real concern of real programmers, even if
> the language is "declarative").

It is interesting, the Haskell prelude do***entation makes complexity
assertions about operations in (for example) Data.Map.  If the prelude
docs also say that some specific pure Haskell spelling of some
operation is supposed to have some specific time and space bounds,
that in turn is an obligation of the language, assuming that the
prelude spec is part of the overall Haskell spec.  

So it seems to me that the language spec should also go ahead and make
requirements about TCO and whatever else it needs to.