Comment by swatow | Hacker News Reader

Comment by swatow | Hacker News Reader

Funny, I once asked the same question about pure mathematics, and got similar answers. In the end I decided pure math wasn't the best bang for buck in terms of benefit to humanity, and decided to work on other things.

When it comes to Haskell, there is a big problem that the people best qualified to speak on the matter have invested a lot of time in Haskell, and therefore are subject to biases associated with their investment in the field.

For one thing, I would ask what programs use Haskell. The only major project I'm aware of is xmonad. This should tell you something about how practical Haskell is.

On the other hand, you may learn something from Haskell. I studied Haskell because I believe there are much better languages out there, and in order to find them we need to understand and improve on the theoretically most advanced languages (i.e. dependently typed languages) and learning Haskell was the best starting point. If this is not your goal, I actually doubt another 6 months of Haskell will help your day to day programming.

Also, what have you managed to learn in the past 6 months? And how were you studying? I found Learn you a Haskell to be the best introduction.

I am confused. If you feel there are much better languages than Haskell, how would Haskell be the "theoretically most advanced language"?

As far as what I have learned and how I have been studying, I did try "Learn You a Haskell" but found its explanations inadequate and sometimes confusing. Searching the web, I find a number of people feeling the same. Meanwhile, I find a large number of people agreeing with your experience the text is the best introduction. I think the fact the supposed best text is utterly failing a sizable number of people also suggests how worthwhile Haskell might be.

> I think the fact the supposed best text is utterly failing a sizable number of people also suggests how worthwhile Haskell might be.

I think it says a lot more about the fact that learning (and, therefore, the best text for learning) is highly subjective than it does about the worth of Haskell.

I mean, how does the logic even work that leads from "some people think that this text is good for learning about X but I find it is 'inadequate and sometimes confusing' to me" to "X isn't generally worthwhile" even work?

> I mean, how does the logic even work that leads from "some people think that this text is good for learning about X but I find it is 'inadequate and sometimes confusing' to me" to "X isn't generally worthwhile" even work?

I'm pretty sure I did not make that particular claim. I have trouble learning from the supposed best text; others have trouble learning from the supposed best text; asking questions gets me treated with hostility and non-helpfulness. I want to know what I am going to get out of learning the language if I put additional time into it. If I am not going to get any more from Haskell beyond (a) what I already know in C/C++ and (b) the ability to put Haskell on a resume, the question of cost/benefit analysis arises quite quickly.

I actually found LYAH to be really good. I picked up on the basics of Haskell very quickly with it.

If you're comparing your progress to some of the more advanced programs out there, you'll certainly get demoralized. But if you're having trouble with just the stuff in LYAH, then maybe Haskell isn't your thing? 6 months is a good long time to spend on learning something, so I'm pretty sure your gut instinct knows what's best for you.

>I am confused. If you feel there are much better languages than Haskell, how would Haskell be the "theoretically most advanced language"?

Let me rephrase: I think that there are much better languages out there, that haven't been invented yet. The most theoretically advanced languages right now are dependently typed languages. Haskell isn't a dependently typed language, but learning Haskell is a necessary stepping stone to learning dependently typed languages.

Regarding LYAH, it's not really possible for me to comment given how little data I have. I know some CS courses for undergrads teach Haskell, I wonder what textbooks (if any) they use?

Reply on news.ycombinator.com