Lars Lundgren <[EMAIL PROTECTED]> writes:
> On 18 Feb 1999, Carl R. Witty wrote:
>
> [Snip]
> > I fear my previous response to this message was a little hasty and
> > overly forceful. Let me try again.
> >
> > I think dependent types are great; I'd love to have the most powerful
> > type system possible. However, I'm also very wary of undecidable type
> > checking.
>
> [major snip]
> So, if i understand you correctly, you think that:
> dependent types are great, BUT
> *maybe* it is to hard to generate good error messages.
> and
> *maybe* typechecking will not work in practice.
> I think that maybe you are overly pessimistic.
Fair enough (although I'd say "cautious", rather than "pessimistic"
:-). I just think we should have a lot more experience with dependent
types in Haskell (by adding it as an experimental feature to one of
the existing Haskell systems, and having people play with it for a
while) before it would be appropriate to push for it going into the
Haskell standard.
I'm curious: how many people have actually written a program in
Cayenne? How many people have written programs that made significant
use of the dependent typing? Has anybody tried to teach a programming
class using Cayenne?
Carl Witty
[EMAIL PROTECTED]