Why GitHub used Haskell for Semantic

- non-nullable types

- immutability as a default

- typeclassses + structs

- abstract data types

- errors-as-values

- monads

Haskell or any other ML language didn't come up with all these things of course, but Haskell is one of the best languages

Take a look at rust -- it's basically got a near haskell-grade type system (and there are lots of ML languages with more flexible type systems than Haskell as well), with C++ like performance. It would have been a lot harder for rust's designers to incorporate such a nice type system without the exploratory work haskell did and continues to do.

All that said, I'm pretty sure Haskell is seeing little adoption because of the learning curve.

> The most successful languages let you be pure-functional where it makes sense and then stateful where it makes sense. Haskell doesn't, really (from my cursory reading about it).

Haskell almost certainly does this, this is actually one of the best features of haskell -- it lets you do functional things and stateful things separately and encourages you to keep them separate.

* Lazy Evaluation -- usually becomes an issue at some point if the project isn't trivial. * Lack of dependent types -- while Haskell's types system is great, this is a missing piece that proves a sore absence for some. It's not very surprising that several functional languages developed after haskell, inspired by haskell, and implemented in haskell (idris, agda, coq) * category theory -- While you can get by plenty well just knowing a few basic abstractions from category theory, it's true that many libraries rely on highly-theorectical concepts, and that a solid portion of haskell development efforts are still tinged with an academic flavor. Just take a look at some of Edward Kmett's coding videos--he typically has one (or several) mathematics/highly theoretical comp sci papers open in one window, while in the other he's implementing a hugely popular library like lens. This sort atmosphere of academism turns a lot of people off. It pegs the language as a theoretical exercise from the get go, and plenty of people are either don't believe in the value of theory or are far too busy to take the time to dive into how it applies to their situation.

I think the academic veneer around haskell is its biggest weakness when it comes to adoption, but that's said, it's still had more success than a lot of other academic, purely functional languages thanks to the efforts of core researchers/users/contributors to break it out of the ivory tower. The ideas are good. Almost every other modern programming language has stolen concepts from haskell at this point--so in some sense even though it's still somewhat niche its influence is practically ubiquitous (just take a look at the list of languages its noted as having influenced on wikipedia! https://en.wikipedia.org/wiki/Haskell_(programming_language)).

In a broader sense, the functional programming paradigm has proven not only that it's a viable one for industry applications, but often superior to object oriented or imperative techniques when it comes to fidelity of expression and preventing bugs. If nothing else, haskell is great because it provides a fairly rigorous environment in which you can explore functional programming concepts, which you can bring with you to almost any problem and any language and manage to derive some benefit.

Some characters are very approachable and easy to play. If you don’t know what you are doing, it’s alright — you can muddle through. If you play them in a really exceptional way, it doesn’t make all that much difference.

Some characters are really tough to play at all; and once you figure it out, they aren’t particularly exceptional and don’t reward further investment.

A “high skill cap” character is one where you can keep learning and learning and your performance at the game will actually get better and better. Some of these characters are also approachable — they have a gentle learning curve. Some of these characters are basically unplayable until you can play them really well — there is an inflection far to the right where you go from dying all the time to actually winning a fair number of matches.

Haskell is like one of these characters. Until you’re really good and know a lot, you’re basically going to ship nothing. This effect is sort of invisible to senior programmers learning Haskell because they are already so skillful and are used to having to skim a CS paper or two, once in awhile, to be able to get their work done. Once you are good at Haskell, vistas really open for you in terms of the kind of programs you can design and build. Many year after setting it aside, I still rely on what I learned about Haskell API design and effects modeling to design reliable, transparent and modular distributed systems (it all starts with the types).

“High skill cap” characters tend to be admired, but not frequently played.