Can someone explain to me why I'm expected to like math and be good at it if I want to write software? How does writing proofs make me a better software developer? I find this kind of thing extremely mind numbingly tedious and boring yet for some reason every software engineer is expected to learn it if they want a degree. Makes absolutely no sense.

This is why software is so shit nowadays.

It's actually incorrect. You need above average math/logic/reasoning skills to be a good SWE but not elite. There is barely any math done while developing. The hidden X factor is linguistic skills and ability as you will be doing a lot of translation and code reading, and will be building your own code structures that need to be easily legible and intuitively constructed. The best SWEs are generalists moreso than aggressive math autists.

>>105620686

Yes, because we force American grads to go through 5 years of this, while Indians can get their piece of Java University shitpaper after 4 weeks and only need to be paid $10k/year to cover it

If American CS education was realistic and cut out the nonsense (while still being rigorous enough to fail out the losers) we could do it in 2 years for far less money and accept lower wages as a result

>>105620699
least retarded nodev

>>105620686
Not an argument. Writing good software is about solving problems in the domain of programming and computers. College courses should be focused solely on programming and learning how computers work at every level from top to bottom, not theoretical math bullshit from the 16th century that has literally zero relevance to programming. Having a class where you have to write an entire software application from the ground up in nothing but x86 assembly would be a million times more valuable than any math class.

>>105620672 (OP)
A lot of real software engineering reduce to a graph theory and optimization problem. There is only so much you can do by duct-taping libraries and buying a solver without knowing some math.

>>105620699
Most people already have those reasoning skills though. Making you write down tedious proofs of problems that you can already solve intuitively in your head doesn't make you any better at logic and reasoning. It's just a waste of time.

>>105620742

>Most people already have those reasoning skills though

They ABSOLUTELY DO NOT. There are endless ways you could proxy real software engineering in coursework and theoretical math is not it, students don't need anything beyond calculus and algebra.

Rules #2 of /dpt/: Programming is one of the most difficult branches of applied mathematics; the poorer mathematicians had better remain pure mathematicians.

>>105620704
I've always wondered why CS is even still the main degree for software engineering. Why hasn't it been replaced by a more practical degree just called "software engineering" that only teaches you how to write software without wasting half the time on useless theory bullshit? From what I hear, most people who graduate with CS degrees don't even know how to write software because they had so few actual programming classes.

>>105620765
Why do you need to proxy it at all? Just have classes where you write actual software. Make one class where you have to write your own operating system from the ground up. Another one where you write a device driver from scratch. Maybe one where you write an emulator for some old computer system. The entire 4 years of a software engineering degree should be nothing but programming every day.
>calculus
Also completely irrelevant to software engineering.

>>105620765
statistics are useful as well

>>105620781
There is also a software engineer degree but Google wanted cs graduates only for many years. At one moment they wanted only PhD.

>>105620817

Because there are specific angles of SWE you need to focus on to build those skills. Writing software combines all of them at once. These days for instance good practical architecture, design, are sorely neglected, you need to hit students with various case studies so that they can think more dynamically.

>>105620822

Stats are without a doubt the most useful math there is, applies to virtually everything

>>105620672 (OP)
Writing math proofs would help you write proofs of correctness for a program using a proof assistant, like Isabelle or Rocq (which used to be Coq). Or informal proofs of correctness. You basically do really informal proofs by induction to yourself when you envision how and why e.g. a while-loop or a recursive function will eventually terminate and that the desired postcondition will be fulfilled after it terminates. They're really informal in that they're just visual, in the same way this picture is a really informal proof that the sum of a series of cubes (1^3 + 2^3 + ... + n^3) is the same as the square of the sum of a series (1 + 2 + ... + n)^2.
That said, most programming is imperative, so I think leaning stuff like Hoare logic would help more than just pure math proofs.

>>105620872
>You basically do really informal proofs by induction to yourself when you envision how and why e.g. a while-loop or a recursive function will eventually terminate and that the desired postcondition will be fulfilled after it terminates.
Yeah and I do it all instantly in my head. I'm starting to think that writing proofs and doing all this format logic shit is for retards who aren't able to reason in their head so they have to spend 10x as much time writing everything down to figure it out. To me that's just mind numbingly boring and makes me want to shoot myself.

>>105620872
but i want to actually finish a program and not masturbate to how mathematically correct it is

>>105620672 (OP)
>Can someone explain to me why I'm expected to like math and be good at it if I want to write software?
when a person is young and they dont have a firm grasp on thing, they tend to flock to math because they like the certainty of it. they might not be able to throw a ball, or paint a pretty picture, but if they turn the crank correctly on a math problem, the result is the same every time. so math typically attracks weak people without a firm grasp on the world, not the football players, and so to do computers attract the same kind of personality.

>>105620904
Wrong. Learn Haskell, finish nothing of note. After 25 years lord your provably perfect fib(x) calculator over everyone else.

>>105620903
People make mistakes. Writing your reasoning out, even informally, helps you catch them and lets the next person understand. And writing it out formally using some proof assistant language will let you be completely sure your logic is correct.
Here's an example where formal verification caught a mistake and OpenJDK's sorting algorithm: https://jurriaan.creativecode.org/wp-content/uploads/2013/04/sorting.pdf
>>105620904
It's really rare to write a whole program and formally verify every part of it. It's usually just the critical parts, if it's done at all.

>>105620929
I've always hated math but loved programming. One is wasting time on theoretical problems that produce nothing tangible in the end, the other is a practical skill that actually produces an end product you can use and even sell to people.

>>105620949
>and OpenJDK's
*in

>>105620672 (OP)
Because software is a large field and contains various computationally complex math heavy problems that aren't properly solved by slapping together shitty libraries like a retard. Also, of all math that you have to learn in a cs degree, induction proofs are the most similar to iterative programming anyway, so stop your whining, you sound like those hs dropouts that are still sore about this impractical concept called differentiation that just came up when they threw in the towel for whatever retarded reason and now earn minimum wage for the rest of their lifes.

Programming is problem solving, and math is nothing but formalized logic. If you fail at that, you simply cannot be a good programmer. You don't need to remember the results, but you should have learned how to solve such problems.

im a literal mongoloid retard with math who struggles with the most basic things and I make a great living as a programmer
Any time math comes up in a work situation I just go "ahh, I'll need to focus up for this one, I'll get back to you when it's done" then get AI to do everything
I'm jeet tier, but I didn't ask to be born a retard. I'm just making do.

>>105620953

Same. There are two types of guys in tech: those who love the process, and those who love the outcome. "Process guys" are adjacent to math hardos, they live for the means and will agonize over it. The other group want to build something cool, useful, use it, share it, and they are oftentimes held hostage by the former group. I wish my education had been more practical instead of 20 math courses, CS became fun after I got through the first 2 years of that and more into the applications.

>>105621277

The math/optimization is a settled science, which is why we use languages like Go/C#/Java/C++ now. They are self-optimizing and math is going to rarely be a concern or bottleneck. Asymptotic analysis IS useful however, the key thing here is that there are specific forms of math that are useful and complementary and should be included as part of a more applicable, practical curriculum. Rather than "Calculus 7" you could have "Optimizations" which teaches you how to conduct asymptotic analysis, find bottlenecks, fix bottlenecks, common strategies.

>>105620672 (OP)
Programming is essentially discrete math and linear algebra.

>>105620672 (OP)
>I find this kind of thing extremely mind numbingly tedious and boring
Then by the Curry-Howard isomorphism you also find programming mind numbingly tedious and boring and shouldn't bother with it.

>>105620672 (OP)
Maths is a good filter to avoid getting too many retards in programming roles.

>>105620672 (OP)
>posts easy math to make his claim
Come on, bro. This is not difficult whatsoever. It takes almost no abstract thought.

All the math you need for even "advanved programming" outside of research or extreme optimization is not difficult.

these are filter classes one, and two what you're doing here is essentially what programming is. if you don't enjoy the constructing part then you're probably in the wrong program. It's like a mechanical engineer saying he really likes airplanes, but has no interest whatsoever in fluid dynamics, he just wants to larp as howard hughes.

>>105621974
Herp derp. You need to be able to write mathematical proofs to write software professionally. Durrrp durrr, this will ensure the quality of software. But somehow despite being written by these maths graduates Gmail still takes a fuckload of cpu and memory just to load up an email inbox, it's almost as if these filter classes aren't doing shit, or maybe it's even counteractive and punishing & filtering the people who might actually do a decent job

>>105620781
because change is hard and it would take budget away from decrepit CS departments at stanford and MIT(harsh but at least they've been trying). all you would have to do really is add a new more applied curriculum with some math and CS courses(data & algos, OS, the basics), and you'd have a ready to work graduate. shopify partnered with some colleges in leafland on something called a dev degree. you spend 4 years straight working and studying, while being paid, and end the degree ahead with no student debt.

>>105620872
Why do i feel like this is written by ai kek

>>105620672 (OP)
It's not supposed to make you a better developer it's supposed to squeeze you for more money for the university.
And I do a lot of maths for programming work cause I write custom algorithms.

>>105620672 (OP)

>>105620699
do you think the ability to do simple proofs is "elite"?

>>105620686
fpbp. this post goes very hard.

>>105622126
Or maybe the Google employees who work on Gmail don't get paid to optimize for software quality and efficiency, but rather Google's profit.

>>105622126
t. brainlet that got filtered

>>105620672 (OP)
They are the same thing

>>105622422
I hope you didn't fall for the O(xyz) meme.

>>105622383
I know it's not written by AI because I wrote it. And ChatGPT and other LLMs have a very cadence. Yoy might just not he as good at detecting LLM use as you think you are.

>>105620672 (OP)
I think I am getting better at math finally. This is one of the first math pics I can read and understand most of it. Cool stuff, need more. Also, I always notice there seems to be a split between people who only want to know exactly what they need to know to do they job, and other people who can appreciate and apply what they have learned in one context to a different one.

>>105624014
Oh really? Can you solve this? It's pure math.

>>105624131
do your own homework faggot

>>105624151
>leetcode
>homework
meds

>>105620672 (OP)
in college i took cs classes because i had to. i took math classes because i wanted to. i firmly believe there is no area of study more beautiful and awe-inspiring than mathematics. i was a shut-in, antisocial /pol/tard throughout high school and i genuinely think studying math helped me grow as a person. i began to see beauty in the world and in other people that i had been blind to before. so much of the hatred and resentment i held in my heart melted away when i realized that there was grace in everything.
i always felt bad for the cs and econ students who were just there to check a box. i always wished i could help them to see what i see, but such is the human condition. i can't help you like something you don't like, but what i will say is this: you have much, much more control over how you feel than you realize. what you see as tedium vs satisfaction is entirely a matter of perspective. i know "you just don't have the right mindset, bro" can feel like a copout piece of advice but it's true. YOU create your worldview. this goes for every aspect of your life, no part of you is set in stone.

>>105624176

>>105620672 (OP)
learn math only when Numpy doesn't work

>>105624803
so... always?

I can't help you, but I can tell you that I was a mathematician before moving over to software, and it made learning software trivial.

>>105624131
from functools import lru_cache

p = 10**9 + 7

@lru_cache(None)
def mod_pow(a, b):
r = 1
a %= p
while b:
if b % 2:
r = r * a % p
a = a * a % p
b //= 2
return r

def count_sequences(n, m, k):
r = 1
for j in range(1, k+1):
r = r * (n-j) % p
r = r * mod_pow(j, p-2) % p
r = r * m % p
r = r * mod_pow(m-1, n-k-1) % p
return r

print(count_sequences(3, 2, 1))
print(count_sequences(4, 2, 2))
print(count_sequences(5, 2, 0))

>>105624176
Also I agree with this very much. Nice post anon.

>>105620672 (OP)
Programming explicitly is about math, formal logic, and control theory. My advice to a beginner is to work on projects that interest you and learn the math you need as it becomes necessary. I wanted to do some audio software so I learned a lot about wave and signal theory. Then I wanted to do a game so I learned a lot about 3D geometry and some calculus. I think this actually helps me understand it better as I immediately apply the math knowledge and iterate my understanding of the concepts rapidly and by testing them. Even getting weird errors can help you understand the math more as you fix them.
If you think your programming doesn't involve math it's just that you are calling math libraries other people have written and at some point you'll bump into that reality. Even a GUI library that pops up a text box is doing math to render the text and place everything where it should go.

>>105625219
>Even a GUI library that pops up a text box is doing math to render the text and place everything where it should go.
That's pushing it. Maybe there is a bit of math in font rendering with Bezier curves and whatnot, but if you're talking about the relaitve placement of widgets, this doesn't count are more than doing stuff on an array is "doing math".

>>105625624
>this doesn't count are more than
*any more than

>>105621974
>what you're doing here is essentially what programming is. if you don't enjoy the constructing part then you're probably in the wrong program.
Mathlet cope aside, this is a faggot argument. Real programming is a lot more engaging than abstract mathematics. I didn't stop hating math until after I learned to program.

>>105620672 (OP)
It's absolutely useless information that you won't need in 99% of jobs and ultimately it's something you self-select into. This math is important for people writing core language features, e.g. you're writing the core sort() function everyone else uses. And unless you are writing a game engine or core library functions for machine learning you don't need to know more than basic 6th grade algebra.

>>105620672 (OP)
You're a nigger. Just prove the hypothesis and then show how n -> n+1 you fucking retard NIGGER. You have infinite n to work with actually and you are a retard and basically fuck you gorrila NIGGER brain dead IDIOT

>>105624176
Based. There's a harmony and beauty underlying all creation. The greatest value is harmony, above all else. And nothing you do or endeavor to do can ever be disconcordant to the silent music that is always playing

>>105620672 (OP)
Korean students do this in their high school

>>105625926

Korea is a degenerate country that has retarded laws and standards. I wouldn't use them as an example for anyone anywhere else

>>105625926
Name one thing Koreans have invented in the modern era.

>>105620672 (OP)
I mean, I've never found this demanding. There's no hard part in the logic

>>105620686
more mathematicians are programmers than ever
they're also just as bad at it as javascript bootcamp ninjas

>>105625942
>>105625981
Koreans don't have GNU spirits; their movements all come down to money

The only open-source project that's I know made by a Korean is this
https://github.com/swc-project/swc?tab=readme-ov-file

In a world where programming languages exist, math syntax should be hated and considered deprecated notation.
Nobody is compiling that shit for any practical purpose, it's only useful for academia in an era where you're limited to writing finite symbols on small slabs, not one where whole real programs can be expressed in something that looks almost like plain English.
Ancient Greeks do not @ me.

>>105626051
trvke
so many of my cs professors made relatively simple stuff look like hieroglyphics by using their stupid fancy math notation instead of pseudocode

>>105620672 (OP)
>Can someone explain to me why I'm expected to like math and be good at it if I want to write software?

Proxy IQ test, we don't want dumbfucks like you in the industry :)

>>105625624
idk man, my game physics programming was still just "doing stuff to arrays."

>>105625219
Great post. Brute forcing theory before practice is inefficient and leads easily to burnout. Good to have some foresight into what's out there and available though.

Way more fun to study along the way and stuff actually "clicks" when it solves a real and immediate problem.

>>105623649
I don't spend hours calculating it if that's what you mean. But obviously I avoid quadratic complexity. Most of the time that just means not putting things into nested loops.