Haskell can be very readable and elegant, but what I do find frustrating is when research papers show sample Haskell, but it's not quite Haskell. They just can't seem to resist the urge to replace various operators and types with their own non-alphanumeric symbols, thus making what should be reading simple code into some kind of hieroglyphic translation exercise..
For the most part the code in the research papers actually _is_ valid Haskell, provided you enable the -XUnicodeSyntax extension and include a few extra modules. Haskell 98 permits Unicode symbols and identifiers; the extension covers the handful of symbols which are built in to the language or enabled by other extensions. You can find a full list here: https://downloads.haskell.org/~ghc/latest/docs/html/users_gu...
Personally I prefer Unicode symbols and tend to use them whenever possible. I have some XCompose rules configured to make them easier to type. Unfortunately, I have yet to find a comparable input system for Windows.