> Sometimes I wonder if some language communities really are cults. It's the same with Lisp - replace "type system" with "homoiconic" and you have the exact same mantras about ultimate supremacy and "intellectual" superiority. But the proof of the pudding is in the eating and across probably 10 computing devices in my home, not a single one has any Haskell or Lisp programs running on it. Lots of Java, C, C++, Python, Rust, Bash, JavaScript, even Perl (probably) but not a single binary compiled from Haskell or Lisp interpreter to be found. To any reasonable person, that should be a strong signal about the so called "power" and value of their sacred cow.
Xmonad and pandoc are probably the most heavily used Haskell programs. Not necessarily contradicting your overall point, just exhibiting the two examples that someone might reasonably be expected to run into.
I got rid of xmonad in favor of i3 ages ago - precisely because it doesn't force me to edit a Haskell script just to change a keybinding. Obviously just my personal anecdote, but this 'feature' alone probably accounts for a significant % of users either jumping ship, or never boarding it in the first place.
fair enough - i don't use xmonad but i do use pandoc on occasion. in full transparency: it occurred to me that i am familiar with postgrest, though i've never run it personally. still no lisp programs though!
Xmonad and pandoc are probably the most heavily used Haskell programs. Not necessarily contradicting your overall point, just exhibiting the two examples that someone might reasonably be expected to run into.