Comment by ruslan_talpa | Hacker News Reader

Comment by ruslan_talpa | Hacker News Reader

Parent Full thread

I am not sure on which side of the flame war you are on :)

How is being more distant from pure math result in things being more expressive (Java more expressive then Haskell)

People will use whatever is more familiar (lazy), no one really wants to (put in the work and) learn something new. All of us have been learning the imperative way since childhood, that's why each new imperative programming language can be picked up in a week, but then we turn and say Haskell is hard and forget to consider that we've invested a lifetime in learning things like C/Java and gave FP a few weeks.

Hardware does not need to change at all, Haskell works just fine as is.

Complex software in Haskell is hard to read only for ppl that never got beyond "Learn You a Haskell" (i am also not that far along) (but i'll admit there is a tendency among Haskell developers to be more cleaver then it's needed with their code and unnecessarily cryptic). The solution to a complex problem is 100 time easier to read in Haskell then in any other language.

I stay away from flame wars and actively attempt to see the whole picture.

My experience is my own, viewing students confront different languages and just looking at it all grow. I'm biased towards myself, heh.

I certainly agree that haskell is awesome. And I know what you're Trying to say.. We're Sadly educated to prefer a certain syntax, and many programmers are still bad at math, bar a devoted and lucky core.

I think it's a pity that whenever someone dives into haskell, in the first few tutorials/pages he immediately is show how this great beauty comes from math and category theory and how it's all connected and they get discouraged very fast (if they were not exposed too much to college level math). If you find the strength to ignore that for a bit in the beginning and find a path laid out by a professional developer, you can do amazing useful stuff without understanding the math behind it that would be quite hard to accomplish in other languages. Anyway, whoever thinks at diving into haskell, ignore everyone that says it's hard, don't look at math staff in the beginning, and don't try to do to much IO like reading/writing from files/db all over the place.

> don't look at math stuff in the beginning

It's inherently more math-y: No state, just functions that have a given output for a given input. Ignoring things that I would call "math stuff" means ignoring the language, beyond trivial expression evaluation.

> and don't try to do to much IO like reading/writing from files/db all over the place.

Ah, cool. Just don't do too much of the useful stuff, and you're good.

Haskell, and FP in general, has always seemed to me like something that you just have to bite the bullet and dig into. Take it for what it is, and not what someone else tells you about it.

In defence of my advice with which you seem to disagree.

Avoid math stuff: By math stuff i mean "category theory", you don't need it to do useful stuff and it's better to avoid it when you learn the language and get comfortable with it. You also seem to imply that (pure) functions are somehow math-y, i am not sure most of the developers see a function definition and think math.

Don't do to much IO: I didn't say "any", i said too much. It's a way to avoid getting thrown at the deep end of the pool (monad transformers) when taking your first steps. The two most popular Haskell tools/projects, Pandoc and PostgREST have very little IO in them, and yet they are immensely useful.

About biting the bullet: I think expressions like these propagate the notion that somehow haskell is harder then other languages and you just have to expect that and get used to it, thus ppl don't even bother an run away when they hear Haskell. In my personal experience (following the 2 things above), after 2-3 months of learning Haskell, i was able to do this https://github.com/ruslantalpa/skin which 1 month later became the base for the current PostgREST core. Sure the was refactoring and such (with the help of more experienced haskell devs) but the fact is, a few months into the language i was able propose and implement deep changes to useful Haskell tools. My ability to do those things did not come from me being some kind of genius (I mostly did PHP and JS all my career) but precisely because Haskell gives you such powerful tools and is easy once you give it more then 2 weeks of your time.

What was your path to learning Haskell? Why do you feel like you had to "bite the bullet"?

Unfortunately, to do much that's actually useful, you hit IO real fast. That was the most difficult thing that I remember from my undergrad intro to programming class in Haskell.

exactly why we need a change on the level of school curricula - and the lazy part of being human can't be avoided. it's that "maximal result/minimum effort YOLO" that's coded into our genes.

Reply on news.ycombinator.com