Thursday, October 22, 2015

Catanastrophe: The Catagorical Mistake

I'll admit it, I am pleased with myself. I predicted on LtU that Haskell's fixation on category theory would lead nowhere, and some people now seem to agree.

The problem: software engineering concentrates on analyzing and describing rich (algebraic) structures whereas category theory concentrates on the trivial, the superficial, structure of theories. There's nothing there to lift, except for the trivial, which you probably could code in half an hour anyway. So you end up with incomprehensible theories on elaborate types providing all kinds of annotations describing structure and functionality you'll probably never use, where you could code the functionality you need yourself, but are forced to comprehend all the trivial observations anyway.

Of course, Haskell is a research language, a moving target, and I've no problem with being proven wrong somewhere in the future. But for the moment, I'll assume I was right and now other people start to find out and agree on that.

So maybe it's not such a bad idea to have an untyped functional language?

No comments:

Post a Comment