Thread 16824912

Why don't mathematicians use this language? All I ever see is MATLAB or R. Sometimes Lean or Coq which are respectable. But never Haskell, despite being the premier (non-proofs) language for mathematicians, designed by mathematicians.

>>16824912 (OP)
Because most of the time someone doing something mathematical with code is working with vectors and matrices, and will be doing a lot of numerical work. MATLAB, Mathematica, R, Python are all better for that than Haskell.

>>16824914
Are they?
Isn't the ideal for that stuff Haskell given how clean and breezy point-free style is? You just use a zipper to apply a large composed function to a 2d array/vector, or a (Float * Float) for vectors/points

Oh wait it already has a matrix monad built-in
https://hackage-content.haskell.org/package/matrix-0.3.6.4/docs/Data-Matrix.html

>>16824912 (OP)
>All I ever see is MATLAB or R.
Kek. Those are statisticians.
Real mathematicians use Fortran77.

>>16824912 (OP)
There was OCAML before

>>16824920
Why not Haskell though? Legacy code?

>>16824912 (OP)
It was designed by CS monkeys cargo culting math they didn't understand, not mathematicians

>>16824942
Why do mathematicians hate category theory, type theory and lambda calculus so much?

>>16824922
>Why not Haskell though?
All I need is a Gödel number and a UTM.
And a lot of scratch paper.
>Legacy code?
"Sandbox" has an entirely different meaning in maths, Codie.

>>16824970
You didn't answer my question.

>>16824912 (OP)
Mathematicians don't use any programming language.
Mathematicians sit in a circle jerking each other off of how superior they are, while begging for scraps on the sidewalk.

People who use programming languages are people who use mathematics to do useful stuff, not mathematicians.

>>16824978
I think mathematicians are useful, and their theories are cute. Prime numbers and all of that jazz.

>>16824977
Answered both, but feel free to ask follow up questions, Whinoceros.

>>16824986
You're answers are incorrect and incomplete.

>>16825021
>You're answers are incorrect
Impossible
>and incomplete
That's your understanding. Ask your follow-up questions or stfu.

>>16824912 (OP)
Who said they don't

>>16825102
Me. I did. The mathematician in the thread.

A

>>16825104
Well then, doesnt it depend on the field?

>>16825130
Math v. Science thread
>depends on the field
I shit you not. Funniest fucking line I have ever read. A new Hughe Durnahm, I tell you.
Still lol'ing.
My sides.
Ouch.

>>16824912 (OP)
>all i ever see is MATLAB or R
so you're not really doing work where you should use haskell in the first place

>>16824917
I mean it really depends. If you're doing research then it makes sense that you'd reach for the easiest tool to use, and matlab/python have incredible amounts of support available. Remember the goal is to finish the research so you can get funding for the next project, not to produce clean code. Clean code is the job of the engineers who use your research, if it turns out to be useful.

>>16825199
I find that Haskell is massively preferable to Python for research. It's more expressive, not prone to annoying runtime errors thanks to the type system, and there are a lot of really powerful extensions for the compiler. Linear programming, full blown proof system, etc. All just in your type system alone.

Python is fine for small programs, but it doesn't really have anything on Haskell here. Even when it comes to third party libraries, 99.9% chance Haskell has a counterpart and it probably works better there.

I think it really comes down to familiarity. If you already know python, which is likely, then you save yourself the trouble of learning Haskell. But I think the upfront cost is worth it, because the nature of Haskell really does save a massive amount of time when writing non-trivial programs.

>>16824985
Yeah yeah we all suck their dicks to appease them, cause if they got off their high horse and decided to be useful we'd all be out of jobs.

>>16825263
>if i graduate, i will be forever unemployed

>>16825380
if you have published papers in haskell and can't get a job in the finance industry or at facebook or microsoft at the very minimum, you're doing something wrong.

>>16825487
>A->B
I have never had any of those issues.
~A->B
Hence B, regardless. Anything else that Haskell has nothing at all to do with?

>>16825509
you could get a really good job if you advertised you used haskell in your research, so the proposition that using haskell in your research will make you unemployable is necessarily false. if you're concerned with a post academic career, haskell still makes more sense than python.

do you really think you're going to be happy gluing webservers together?

>>16825513
Made over $530K last year with the Gödel numbers last year.
Good luck with your blog!

>>16825519
Your boss made 100x that without touching a keyboard

>>16825575
>your boss
Lmao

whoever invented this language is a fucking idiot it was aids just to do hello world

>>16824912 (OP)
Most mathematicians are not working on programming language theory, so they don't give a shit about category theory or laziness autism.
Maybe you could teach them to use some ML-like language.

>>16825519
Genuinely interested, could you give more details or a paper?

>>16824961
>category theory
What are you on about? Category theory is ubiquitous in math.
>lambda calculus
Most mathematicians don't know a thing about lambda calculus, much less hate it
>type theory
Similar deal as with lambda calculus, but I guess it's more well-known. Most mathematicians just view it as a waste of time to reformulate the entire framework of math, when the one we have is very familiar, and it's completely unclear what, if any, advantages type theory might have. If you ask about homotopy type theory, most mathematicians want to avoid it, because it just seems like a shitshow.

>>16824912 (OP)
It's not how they think. I've had heated discussions with them about Haskell (and declarative languages in general) or theorem provers. They really hate that stuff.
I guess because it forces them to make things explicitly algorithmic with a clear path from axioms to result, which would be intuitionistic/constructive math while they prefer to play word games with assumptions.

>>16826661
Wow. Issues.
Sorry about that gaping wound in your pride. Don't know how it happened, bit you need to get treatment. It's necrotizing.

Semi related: best way to animate physics?

>>16826722
Me again.
I had in mind python and matplotlib. Basically solve some differential equation(s) and store the solutions in a data structure. Then iterate over it to display something at every point, time, etc. Can I do something like this with Haskell?

>>16824978
>>16825276
What languages do you use for making math models? Out of genuine curiosity.

>>16824912 (OP)
they use Idris though, which is based on haskell

Haskell is the fedora of programming languages

>>16824912 (OP)
big fucking matrices have a handful of highly optimized CPU libraries

>>16824912 (OP)
Mathniggers are stupid and coding filters them so they have to use baby languages.

>>16826996
And one of them is massiv, which works even in a distributed computing environment

I miss Shark Week and this thread sucks.
Not a single line of working Haskell code. I've seen better ethereum contracts from feces flinging monkey lawyers.