[Haskell-cafe] Re: Why purely in haskell?

[ajb at spamcop.net]

G'day all.

Quoting jerzy.karczmarczuk at info.unicaen.fr:

> Perhaps another example is more relevant, the tradeoffs
> space-time in the
> "optimized" version of the powerset generator...

Yeah, I was going to bring up that topic.

I've been lazy programming for 15 or so years, and the only "bugs" that I
can think of are:

1. Indirect "black holes" that are not expressible in a strict
language.  You generally have to be doing something bizarre for this
to occur, and it doesn't take too long before you can accurately
predict when they constitute a likely risk.

2. Space leaks.  This is the one thing that really bites hard, because
a space leak is a global property of a lazy program, and hence can't be
reasoned about locally.

Cheers,
Andrew Bromage