A couple of very good Haskell articles I have seen recently:

• Introducing the Monads by Aditya Siram. An great overview which serves to confirm/reinforce things that I have learnt or figured out so far. Tho’ I think uses return where it is unnecessary; for example the first snippet on p.4 should be:

  mport Data.Char
  main :: IO ()
  main = do
       writeFile "test.txt" "a,b,c,d,e"
       x <- readFile "test.txt"
       print (map toUpper x)
  

  IMHO. Into what are we injecting y? Surely if you have an IO String at the first opportunity you would want to ← it to a String so you could play with it in pure code (i.e. toUpper in this case).

  Also I wasn’t aware of this technique, which I will definitely be using:

  import "mtl" Control.Monad.State
  

  Then compile with:

  ghc --make -XPackageImports ...
  

  Since I seem to have both monads-fd and mtl installed. Previously I had been using -hide-package.

• Monad Transformers Step by Step by Martin Grabmüller. Much more theoretical, I think I probably need to read it a few times to get the most out of it.

Normally I learn a language by just dropping myself in at the deep end and saying right, this program that I need to write anyway, instead of using (language I know) I’ll do it in (new language), God help me. Pascal, MATLAB, FORTRAN, C, C++, Perl, Java, Python, Tcl/Tk… I started in Python using it as I would have used Perl (yuk! but I was young), then before long worked up to HOFs, generators, decorators, etc. Same with Tcl. My early Java looked like C and my early C looked like Pascal.

But that doesn’t work for Haskell or OCaml. I tried it before and had to give up, this time I am being much more methodical about it, getting the foundations in place, and it’s paying off so far.