[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: antipathy to type systems



Olivier Lefevre wrote:

> [type errors]

> OTOH, as Shriram pointed out later, there is plenty more that go wrong
> even in a well-typed system: think concurrency issues, for instance.
> Yet some type advocay has a cargo cult feel to it, like types were the
> be-all and end-all of program checking.

You bet.  To those who were present at LL3, you'll recall that
Matthias and I raised the issues of type errors to the speaker who
said something to the effect of, "The Haskell type system just takes
care of it -- you don't even even need to know about Haskell's types".
Needless to say, I think that claim is at best semi-true -- it's true
until you encounter the first type nasty error.

That said, let me reiterate my first post.  These critics would be far
more credible if they could post interesting examples of what they had
tried to do in ML but failed to accomplish because the type system got
in their way.  (I have a few examples, but then I'm pretty sure I fall
in category four.)  Until they do, there won't be any *dia*log.  Each
party will keep talking to and reinforcing its own constituency.

Shriram