Great article Vincent, I related so well to what you were talking here despite not doing Haskell programming but Julia programming. Just like you are in love with Haskell I think Julia rocks but it has many similar problems.

Julia is dominates in academia rather than in industry. It is a bit different from Haskell though in that it is not language and type theory academics which dominate but scientists within physics, chemistry, astronomy, climate change, economics or pretty much whatever you can imagine.

So the challenge is a bit different. Julia is probably a lot easier to learn than Haskell, but the community is still very dominated by very smart people. Going to JuliaCon is a great way of feeling like an idiot. I did actually have quite good grade in mathematics at Uni but I am totally out of my league around lot of the people who use Julia. Mathematics is kind of like their mother tongue and that is reflected in a lot of the code.

I really believe in Julia as a general purpose language for all people. I have made a video course, self published a book and later created a book with Manning that is about to get published. So I am serious about trying to help make Julia a broader language. But it is challenging. Many don't try it because they think Julia is only for science.

Like you guys we also have a documentation problem. One of the issues we have is that Julia is extremely composable but people are used to big monolith libraries in other libraries documented all in one place. Julia however is so flexible you rarely need big monolith libraries as it is so easy to combine almost any library. That is very hard to convey though.

People in the Julia community still don't really know how to teach that or make people understand the concept. People will lookup a Machine Learning library in Julia and decide they cannot do ML in Julia because it looks too small to be done. But it actually IS done. It just doesn't need to be big.

I think Haskell has some of the same issues. Powerful languages are so different from run-of-the-mill languages that it is hard to convey to people just how powerful they are. And you cannot really explain why in 3 lines.

I have actually contemplated learning Haskell lately. I wrote about Unison in various articles. It is heavily inspired by Haskell and implemented in Haskell which means it has many of the same core libraries as Haskell.

https://erikexplores.substack.com/p/writing-code-without-plain-text-files

https://erikexplores.substack.com/s/languages

It was sort of like a gentle intro to Haskell. I have studied Haskell before but didn't quite have a reason to spend all the time. With Unison the motivation was to learn non-file based development and algebraic effects. I honestly didn't like the language much at first. Never been much of a Haskell fan.

Yet after suffering through for about a week, it started to really grow on me and now I started thinking it sucks that it is really an experimental language with no v1.0 release yet. That got me thinking about Haskell.

But I don't really have a good reason to learn it. Only reason might be to have an alternative to Julia when I need to produce small native binaries to distribute.

Ironically I think Julia paved the way for me. I do a lot of functional style programming in Julia so a lot of the Unison concepts were familiar to me like: partial application, function chaining and function composition, filter, fold, map and all that.