I am going to update this wiki as I am learning Haskell, and reading your comments.

Learning Haskell was quite hard for me, and I needed to be confident that my own code really works, and under which conditions. This is may be the same for you.

The reasons for the difficulties in learning Haskell, may be:

• Before I started with Haskell, I was only used to imperative languages (Assembler, C, Pascal, Modula-2, C++, PHP, and more). The most, I was used to C++.
• The difference between the imperative paradigm and the functional paradigm is like the difference between hockey and chess. It is both sport, isn't it?
• Often, the code examples are not appropriate.
  • ..or I am to stupid?
  • A lot of the available examples are outdated, and incompatible with current packages, libraries and compiler versions.
  • Often just incomplete.
  • Often they contain confusing useless, warning generating statements.
    • Example here: Confusing pop statement in LYAH
• Some of the available documentation is incomplete or inaccurate.
  • See the comments in here
• Some information are only available on stackoverflow or reddit.
• If you place questions on stackoverflow, and don't know how to formulate the question from the very beginning then you are also risking your reputation, like here.

I consider myself a techie since I was a curious child. At an age of approximately 8 years, I assambled an electric DC Motor, and believe still to remember that I was shocked and amazed that it worked. I was so keen to understand…

This craving to understand is a very big driver - for me.

Just to imagine how rich Haskell is in terms of abstraction - and to understand how must can be and sometimes has to be understood - enyoy the example below.

The following shows how 10 primes are evaluated and displayed starting with the one millionth prime number.

• Example )¹ )²:

  • module Main where
     
    import qualified Data.List as L
     
    main :: IO ()
    main = do print ((genericTakeFrom 1000000 10 primes) :: [Integer])
     
    genericTakeFrom :: Integral a => a -> a -> [a] -> [a]
    genericTakeFrom nFrom nCount l = L.genericTake nCount ((L.genericDrop nFrom) l)
     
    primes :: Integral a => [a]
    primes = 2 : [n | n <- [3, 5..], all ((> 0).rem n) [3, 5..floor.sqrt$((fromIntegral n) :: Double)]]

• Output:

  • [15485867,15485917,15485927,15485933,15485941,15485959,15485989,15485993,15486013,15486041]

To understand this code (8 lines of code) completely I have to understand the following:

1. Packets
2. Import
3. Prelude
4. Monads
5. do-notation
6. Type inference?
7. Types
8. Type signatures
9. Type defaulting
10. Type classes
11. Type constraints
12. Lists
13. List comprehension
14. Recursion
15. Function composition
16. Lazy evaluation

And there is also

• design,
• optimisation,
• test, and
• proove of validity

• Haskell is fast.
  • The execution of the algorithm above takes approximately 1 minute on a modern computer (Year 2021).
• Haskell supports parallel execution.
• Is very good testable and validatable to rock solid code.
• Haskell programs can solve problems that I would not dare to solve without - at least without a functional programming language like Haskell.

)¹ The code may not work with extrem large numbers. I do not know at which conditions the floor function would fail. My calculator says sqrt((1E16x1E16)-1) is 9.999.999.999.999.999,9999999999999999, but sqrt((1E17x1E17)-1) is (1E17)!

The following code (sieve of Erastothenes) will also work but by far not as fast for large numbers (approximately 100 times slower for the first 20 thousand prime numbers):

primes = let sieve (n0:lrn) = n0 : sieve [ n | n <- lrn, n `mod` n0 /= 0 ] in sieve [2..]

)² The code compiles without any warnings with ghc 8.10.4 and option -Wall.