Re:How to Learn Haskell
by fusiongyro on Wednesday April 11, 2012 @01:31AM
(#3826084)
It's going to depend a lot on what you're trying to do and what your personality is like.
OCaml is probably the most pragmatic of the three, but the parallelism and concurrency story is pretty weak. There is a book, OCaml for Scientists, which is a tutorial intended for the scientific audience though. It's also heavily used in finance. It also compiles quite fast.
Haskell is the hardest of the three to learn and get up to speed with. If you have a heavy mathematical bend, it may make the most sense. Parallelism and concurrency are simple and easy, but I don't know if you can do MPI or OpenMP style supercomputing; it really seems optimized for desktop/server processors.
I love SML for its simplicity, clarity and power, but it is essentially a dead language at this point, with most of its userbase having migrated to Haskell or OCaml. I don't know if Concurrent ML supports true SMP, but that would leave you limited to certain implementations. MLton is fantastic, but it comes at a price of very slow compiles with no separate compilation.
If I were in your shoes, I'd look at some sample code in OCaml and Haskell and see which one looks more reasonable, and then write a small program in each and see which one you like the feel of more. You're probably going to run into a sticky spot at some point in your project whichever one you choose; the more you like the option you have, the more likely you are to stick it out.
OCaml is probably the most pragmatic of the three, but the parallelism and concurrency story is pretty weak. There is a book, OCaml for Scientists, which is a tutorial intended for the scientific audience though. It's also heavily used in finance. It also compiles quite fast.
Haskell is the hardest of the three to learn and get up to speed with. If you have a heavy mathematical bend, it may make the most sense. Parallelism and concurrency are simple and easy, but I don't know if you can do MPI or OpenMP style supercomputing; it really seems optimized for desktop/server processors.
I love SML for its simplicity, clarity and power, but it is essentially a dead language at this point, with most of its userbase having migrated to Haskell or OCaml. I don't know if Concurrent ML supports true SMP, but that would leave you limited to certain implementations. MLton is fantastic, but it comes at a price of very slow compiles with no separate compilation.
If I were in your shoes, I'd look at some sample code in OCaml and Haskell and see which one looks more reasonable, and then write a small program in each and see which one you like the feel of more. You're probably going to run into a sticky spot at some point in your project whichever one you choose; the more you like the option you have, the more likely you are to stick it out.