I've toyed with Haskell, ultimately moving on to the Ocaml/F# camp. There are only two things that I miss from Haskell without an appropriate equivalent or easy workaround: Type Classes, and Higher Kinded Types.
This big roadblock that everyone claims with Haskell, Monads, didn't give me any problems at all...even if it didn't make any sense to resort to them to do something as trivial as IO. What really turned me off more than anything was the combination of the academic focus of the community combined with this weird culture that I could only describe as a cleverness competition.
I'm a pragmatist, and it is obvious that real software can be created with Haskell, but it doesn't really come through in the tutorials or books. There is a lot of "Look, this is really cool!", but rarely a follow up with "and this is why it matters!". Everything seemed to revolve around cool tricks with no practical concern behind them, and then in the comments you inevitably find comments from other Haskellers claiming to do it just a little bit more cleverly using <<lenses, GADTs, zippers, continuations, or some other obscure abstraction>>.
Ultimately it was Ocaml and F# that taught me all of the really important lessons of the ML family, despite the relative lack of learning resources out there. It was there that I found the benefits of making illegal states unrepresentable, expressive pattern matching, type inference, composition over inheritance, etc.
It is a shame, because as languages and runtimes, they probably are inferior to Haskell. Ocaml does fine with concurrency, but parallelism is a disaster. Its class system, its one major distinction from SML, is mostly considered a code smell. Its stance on operator overloading requires me to keep a table of operators handy, instead of the usual intuitive ones. And F# has the sane operators and parallelism story but doesn't have SML/Ocaml Functors, and is still a Windows-first ecosystem (Mono has gotten a lot better, but its still a kludge).
The "and this is why it matters" is often left to be inferred by the reader. ;) General kinds of reasons include "preventing (whole classes of) mistakes" or "writing even more reusable code" or "making things fast".
I know the kind of thread you mean -- it happened to me last weekend https://www.reddit.com/r/haskell/comments/2tl8v1/making_prop... . But I really appreciated those suggestions about even more clever ways to make my code better, even though I have so far not incorporated any of them.
It helped me learn more about the shape of a space I'm only starting to explore, and since I'm exploring it for real-world reasons, the very abstract and hifaluting suggestions were easy to relate back to real-world concerns. Which is a great way to learn.
And I suspect several of the commenters there were commenting for exactly that reason -- if there's one thing the haskell community excells at, it's noticing when someone is in a receptive state (or coaxing them into one), and helping them learn.
The Windows-first status of F# is changing. There have been major improvements in this recently, most notably, the whole compiler being open sourced and its development moved to github[1], and .NET is going to be available on Linux[2]. This is not reality yet, but it will be during this year.
> This big roadblock that everyone claims with Haskell, Monads, didn't give me any problems at all...even if it didn't make any sense to resort to them to do something as trivial as IO.
That's a bit uncharitable. You're likely going to end up working with Lwt (monadic concurrency/IO library) as soon as you need to do something interesting with IO in OCaml.
> Everything seemed to revolve around cool tricks with no practical concern behind them, and then in the comments you inevitably find comments from other Haskellers claiming to do it just a little bit more cleverly using <<lenses, GADTs, zippers, continuations, or some other obscure abstraction>>.
While I agree that showering code in mathematical abstractions doesn't make for readable code, a zipper is not "cleverness for cleverness' sake". It lets you manipulate easily manipulate an immutable tree (and it doesn't decrease readability).
> Ocaml does fine with concurrency, but parallelism is a disaster.
Hopefully the multicore runtime will turn into something concrete eventually :)
It's also lacking real type classes, at least until modular implicits are ready.
This big roadblock that everyone claims with Haskell, Monads, didn't give me any problems at all...even if it didn't make any sense to resort to them to do something as trivial as IO. What really turned me off more than anything was the combination of the academic focus of the community combined with this weird culture that I could only describe as a cleverness competition.
I'm a pragmatist, and it is obvious that real software can be created with Haskell, but it doesn't really come through in the tutorials or books. There is a lot of "Look, this is really cool!", but rarely a follow up with "and this is why it matters!". Everything seemed to revolve around cool tricks with no practical concern behind them, and then in the comments you inevitably find comments from other Haskellers claiming to do it just a little bit more cleverly using <<lenses, GADTs, zippers, continuations, or some other obscure abstraction>>.
Ultimately it was Ocaml and F# that taught me all of the really important lessons of the ML family, despite the relative lack of learning resources out there. It was there that I found the benefits of making illegal states unrepresentable, expressive pattern matching, type inference, composition over inheritance, etc.
It is a shame, because as languages and runtimes, they probably are inferior to Haskell. Ocaml does fine with concurrency, but parallelism is a disaster. Its class system, its one major distinction from SML, is mostly considered a code smell. Its stance on operator overloading requires me to keep a table of operators handy, instead of the usual intuitive ones. And F# has the sane operators and parallelism story but doesn't have SML/Ocaml Functors, and is still a Windows-first ecosystem (Mono has gotten a lot better, but its still a kludge).