Hacker News Reader

Comment by saagarjha | Hacker News Reader

> Haskell is still stuck with most of the wrong decisions they made in '88

Which ones?

While I don't agree with the general snarkiness towards Haskell of the commenter you are replying to, being lazy by default could be seen as Haskell original mistake. It was interesting from a research point of view though.

Miranda is also lazy, so that can't be it.

it's hard to call the sine qua non of a language a bad design decision. if it were, the language would be dead. haskell survives because it's a lazy purely functional language with some industry support. bad design decisions have to be something apart from that.

I can't disagree more.

What set Haskell apart is how it successfully limited side effect to the IO type and the introduction of type classes.

Laziness by default is definitely a bad design choice as far as I'm concerned. The drawbacks are not worth it.

:: for type signatures, for one; I think that came straight from Miranda.

Honestly though there weren't too many outright "mistakes", but some parts of Haskell have evolved significantly, for example the introduction of the IO monad and all the associated type classes, compared to the original based on infinite lazy lists.

Answering here so everyone in this thread can see:

The most obvious one (and, admittedly, the one that immediately comes to mind at almost 3AM) is their exclusion of non-linear patterns. It didn't reduce complexity, yet made the language worse for the purpose the professor stated in an obvious way.

That reminds me of n+k patterns, another weird wart.

Reply on news.ycombinator.com