/sci/ - Science & Math

> CLRS
> Not TAOCP or Analytic Combinatorics by Sedgewick
> No analysis or algebra or any serious math topics
NGMI

>>12644922
So you've read TAOCP?
Name every instructions for MMIX

Everyday until you read it

>>12644893
stay in >>>/g/

>>12645225
Sorry for ruining your chain of IQ/race correlation threads, faggot.

>>12644922
>TAOCP
>Analytic combinatorics
both great collections / books, but unfortunately not everyone in CS is interested in these topics. If you aren't doing hardline algorithm analysis, the complex analysis in analytic combinatorics is too heavy, and CLRS is largely derived from the most salient introductory parts of TAOCP anyway.

>>12644893
This list https://rentry.co/rkogc
was posted as a compilation of books I had suggested. I'm currently reworking this list into a small guide if anyone's interested in giving suggestions.

>>12645934
is this a good book?

>>12644922
if someone has done most of CLRS they have no reason to bother with TAOCP right? both are introductory and CLRS contains more algorithms and is more up to date.

>>12646294
nevermind apparently most people advise not to use TAOCP as an intro book despite cspg always bringing it up as a better alternative to CLRS.

>>12646358
No one does, it's literally just a single anon every thread

>>12646294
>>12646358
>>12646367
See >>12646051
It's a great book but not what everyone's interests are at. Everyone, however, should do Concrete Mathematics

>>12646051
>but unfortunately not everyone in CS is interested in these topics
Then they need to go back to their /g/erbs and stop pretending to be anything other than lowly trained monkeys

>>12646478
Post your post-doc, github, exercise notebooks, and TC

>>12646478
anon, i like complex analysis, and I think people ought to learn it because it's beautiful, but I can't ask everyone to learn it if their interests have nothing to do with it.
no you're not a codemonkey if you don't want to do it lmao

>>12644893
Reading How to Prove it, SICP, HtDP, Spivak's Calculus, and University Physics. Aiming to create an android to blow my brains out for me.

>>12645130
My programming teacher told us to read this but I haven't done so (and probably never will)

>>12646789
>Spivak's Calculus
ngmi

Anyone here into algos and comp programming

>>12646123
It's the third book in a trilogy.
the first two were something about piss and diarrhea. i read these books when i was 5

>>12647181
Just read Engineering a Compiler instead then

>>12645130
does this benefit any kind of developer? even if I don't do low level programming? I'm just curious. I will probably read it regardless for fun

>>12645130
Nah. As others have suggested, you're probably better off with Torczon's Engineering a Compiler and Pierce's Types and Programming Languages

trying to solve p vs np within the next 6 years. already been working on it for 4 years.

dont give a fuck about anything gay like coding, AI, or anything to do with actual computers.

current textbooks im working on rn.

pic related

i would have gone to school for cs but i taught myself undergrad cs because im not a stupid brainlet, just a regular brainlet.

is there any reason to major in computer science in college? Or should it be learned on the side while studying something like math? assuming you are getting a degree no matter what

I did my bachelors in an IT-adjacent field and kinda veered off into doing IT and development work for a couple of years, and now I want to do a masters in CS. However, I'm lacking all the formal basis in the algorithms and mathematics side of it, other than what I learned through osmosis (not much).

I'll be self-learning it and the OP has good resources, but what do you suppose would convince the applications guys at the university I'm applying at that I'm serious about it? Just saying "btw i'm studying this autonomously" in a motivation letter isn't very convicing, I'm sure. Is there anyway to prove something like this? General admissions tips are also welcome.

>>12648545
Do it for the math portion. CS is not just programming, algorithms and datastructures.

>>12648545
>is there any reason to major in computer science in college?
No.
>Or should it be learned on the side while studying something like math?
Absolutely.

>>12648995
The math in CS is soft and slow. You could do it by yourself in a sliver of the time while going much further than their courses.

>>12649031
Can you tell exactly why do you think that way? In particular, what fields exactly are slow and what do you categorise as "hard and fast" math?

>>12648545
You’re gonna get conflicted answers because of the /sci/ dickwaving and because the difficulty / quality of a CS education varies heavily.
If you go to a school with a good CS program, there’s nothing wrong doing it. The math major is great, but it won’t necessarily get you everything you want unless your interests are pure theory. My personal interests are theory, but it’s hard to tell people who didn’t want that *not* to do EE instead since it’s more adjacent to systems CS. In fact, a more than decent CS program is flexible to your interests with relevant staff and course selection offering challenging material. CS also gets you past way more automatic filters than just a few CS classes like data structures.
On the other hand, there are a lot of really bad CS programs out there that are useless and teach you nothing of value. At that moment, you’ll probably want a degree in something else while teaching yourself the salient parts of CS, which aren’t necessarily related to anything in software. Here, math provides the most consistent base for any anything related to theory, followed by engineering for systems and applications, and physics splits the difference.

People like to say something something self teach CS makes it trivial, but the truth is that with any strong foundation, you can self teach yourself whatever you desire. Math is a safe bet for being decent at most schools, but I’ve also become way harsher toward people talking about the run of the mill math major as anything special - they’re not. Many don’t even take anything beyond baby analysis and baby algebra prior to graduating. Just keep your wits about you and train your bullshit meter.
>t. grad, was math and CS double major

>>12644893
Wtf is computer science plus? Is it like atheism plus?

>>12649512
CS + SWE + IT + Math + others stuff

why is math so hard bros

If I want to self study some math for doing computer science research in algorithms, is list from CS anon's big list a good list to pic from?

- Artin's algebra up to Galois theory
- Stanley or Van-List and Wilson for combinatorics (looks easy but insanely hard)
- Diestel or Bollobas for graph theory
- Enderton's logic for an easier intro, Shoenfield for something more hardcore
- I'd recommend doing something like Abbot's Understanding analysis or baby Rudin if you're serious about theory - the basic forms of reasoning about analytic structures shows up more than you'd imagine. But the same could be said about anything if you care about theory - Erickson himself studies topology in CS.
- If you care at all about graphics, use CMU and Stanford webpages alongside Milnor if you're brave.
- analytic combinatorics

Which one should I do if I can only read one before the summer starts?

https://www.hudsonrivertrading.com/careers/job/?gh_jid=2160225
What sort of CS student do you have to be to land an internship like this?

>>12649708
What list? And if you had to read only a single book and you've already taken calculus, just take a discrete mathematics course. They're designed specifically for CS majors so include most of those topics baked in. If you want, linear algebra would be the second thing to look to. You can do everything else after if you're really serious

>>12649734
Replace course with book obviously

>>12649708
right now I'm thinking analysis, and maybe Rudin since the list author recommends that if "you're serious about theory". I'm not doing graphics and i will not be expected to know graph theory that well so I can eliminate those. analysis should be more important than algebra right?
>>12649734
this list is what I was referring to >>12646076
and yes I am very serious, I have a pretty much guaranteed undergrad research position lined up and am doing graduate algo courses to prepare. I have already taken all the same math that is required for CS majors. I have done calc, diffeq, linear algebra, and discrete math. numerical analysis is elective for CS majors but not for my major so I have to self study.

>>12645130
how accessible is this to beginners? is it like sicp?

>>12649708
following up on this:
pretty sure I'm going with rudin unless another anon says I should do otherwise. I will probably also ask the professor what math he thinks would be most relevant. Rudin seems to have quite a few books with "analysis" in the name. I is "Principles of mathematical analysis" the one that I should read?

>>12650321
The math links in OP seem to agree that Rudin's Principles of Mathematical Analysis is the gold standard for Real Analysis (Metric Space based)

Math for CS:
https://stackoverflow.com/questions/14868480/math-for-computer-science
https://4chan-science.fandom.com/wiki/Mathematics

>>12649734
>just take a discrete mathematics course.
Those suck
>They're designed specifically for CS majors so include most of those topics baked in.
read: they're watered down for brainlets

What field is there an overlap with language and CS?

>>12651408
https://4chan-science.fandom.com/wiki/Computer_Science_and_Engineering#Natural_Language_Processing

>>12649411
what would be a good vs bad cs program?

>>12649411
>If you go to a school with a good CS program,
These do not exist (in the US). If you found your cs program hard, then you're just a brainlet.
>Math is a safe bet for being decent at most schools
Actually math is the subject who's education varies heavily. School that have programs with baby analysis and baby algebra classes are on the shitty side.

>>12646123
The best

Should I start with TAOCP if I already know some Analysis (Real, Functional, Complex)? CLRS seems handwavy at times imo and the exercises are not that challenging for me.

>>12648545
Honestly doesn’t matter too much. Just pick either one or engineering.

>>12649411
I'm around the top 30 mark for best CS schools as ranked by US news.
I can kind of confirm that everything has been easy so far, although I'm not super far into the major yet.
My strategy is to do as many math course substitutions as possible.

My issue is that I don't have a specific plan for what I'm interested in; I don't have a specific goal, but I enjoy everything I learn that's related to CS. I like learning whatever feels relevant in the world; I like solving most/any types of problems if it seems useful. At first I was a physics pre-major until the drag of learning specifics of things like thermodynamics started to bore me and I realize I don't actually like it that much compared to CS which I know will definitely be applicable in my future at some point.

I think the truth is that college is actually kind of a joke when it comes to education. And this is true for any major that doesn't require special lab/resource access (all math, CS, most of the physics curriculum). You're supposed to make connections in college and prove that you can handle the work. I just want to make the most of the forced learning structure while I'm paying for it.

>>12652494
>although I'm not super far into the major yet.
A good CS program should get harder in 3rd and 4th year. first and second years have to be taught how to program so its easier at first, and first and second years should also be learning computer architecture which also is probably the easiest topic in a CS curriculum. your hardest courses will probably be math and physics. after your discrete math course (which most students found hard at my uni) you should have access to real CS courses.

if the first and second years are so easy for you you can do personal projects or take an extra course or self study math or something.

>>12652434
TAOCP and Analytic Combinatorics are probably good if you have familiarity with the classic algorithms. The latter is especially recommended if you're comfortable with complex analysis

>>12651547
>These do not exist (in the US). If you found your cs program hard, then you're just a brainlet.
Nah, I think this is a pretty unfair statement. There are a lot of watered down CS programs, that's for sure. But there's the well documented internet tough guy who says it's trivial and struggles to get through it. The bravado is most commonly thrown at CS since the popular conception of CS in the US is "study whatever is relevant to writing code to solve a problem you're too stupid to do by hand" - without realizing the fact that code aside, solving the problem naively is easy, but doing it clever is one of the major conceits of CS. Hell, even that aside, I don't see why anyone would be a brainlet for struggling with anything related to graphics processing, advanced algo (ie cache oblivious models and online problems), etc.. It sounds like you're someone from a different major who's made a judgement call despite not really knowing anything past the basics.

I suspect that CS will undergo a transformation like EE, which has long since evolved past the domain of electrical devices, and the major requirements will mature as a result. But not every CS major is a mouth breather, and not every CS program is trivial

>>12652494
>>12653042
CS programs unfortunately try for a one-size-fits-all approach where they filter people with data structures and calculus but don't commit hard enough. So the difficulty of a CS major is down mostly to the selection of electives because the core is generally watered down.
This is why the jump from CS undergrad to CS grad can be so huge for people, and why it's obvious who the good CS students are among everyone else.

>>12644893
I think I prefer math over CS, despite me studying/working as a programmer. Right now I'm getting semi-filtered by fourier transforms and asking myself why I'm working through this for fun when I don't really have time for it. Oh well

Bump

>>12654284
Are you comfortable with proof math? You might want to pick up a book on Fourier analysis prior to signals, where they can be really wishy-washy about how everything works.
What's confusing you about Fourier transforms anon?

I got a C in Algorithms as a Junior. I have the option of either putting the letter grade (C) on my transcript, or having the class as simply Passed (which will not affect my GPA).
Which one should I do? I am afraid employers/grad schools will see it as Passed and assume I got a D- or something, but I also do not want them to see a C. I guess I want it to be vague so they could also assume I had a B- that I did not want for an important class such as Algos.

>>12656989
employer may not even look at your GPA, and they certainly aren't going to look at your individual courses so passed would be better if you plan to go down that route assuming your current GPA is > C

>>12656989
>I got a C in Algorithms as a Junior
How? The class is piss easy.

>>12644893
Shouldn't a thread like this be on >>>/g/ ?

>>12657413
/g/ doesn't care about science and the board is way too fast for a science thread to survive the flooding of /v/ermin brand warring over their CPU parts and /pol/tards telling you to use their botnet browser.

>>12657373
Online school was hard for me.

>>12657373
at my school junior level algorithms is known to be the hardest undergrad course

How long to complete teach yourself cs?
Long story short if I don't learn cs in 6 months im fired.

>>12657516
Which parts do you need to learn? algorithms? A programming language? Also, why 6 months?

>>12657516
What does that even mean? What the hell is "learning CS"? Do you just mean learning how to do a little programming?

>>12657545
>>12657524
I just know little bit of programming. I'm in senior year (Mech) and got really lucky in an interview for a software role. But I only write monkey code.
I'm supposed to join in 6 months.
I need to learn everything one learns in four years of cs.

not sure why he deleted his posts but to the ME anon asking how to cram 4 years of CS: you don't need 4 years of CS to program. most of what a CS major does will not make them any better of a programmer. you get good at programming by writing code. at most you might want to do a really basic book on basic algorithms and data structures, preferably a book that shows you how to implement them in code, just so that you have a little bit of familiarity with them before you see them pop up in code. other than that just cood and don't be afraid to google any questions you have about how to do things or how things work and learn as you go.

>>12657516
Programming and Data Structures: 1-3 Months
Digital Logic, Computer Architecture, and Assembly Language: 1-2 Months
Operating Systems: 1 Month
Programing Languages and Compilers: 2-3 Months
Discrete Math and Probability (@ CS' level): 7-12 Days
Algorithms, Automata, Computability Theory, and Complexity Theory: 1-3 Months

6-12 Months

>>12659212
>Programming and Data Structures: 1-3 Months
this is fine
>Digital Logic, Computer Architecture, and Assembly Language: 1-2 Months
not necessary, you don't really _need_ to learn computer architecture. your employer hired an ME, he knows you're not educated on CS. if your job is with a high level language like python you can ignore assembly and such for a very long time. for computer architecture, the most that a working programmer needs to know is how memory access and caches work.
>Operating Systems: 1 Month
again, not necessary, I wouldn't worry about it. I would put basic architecture and assembly first.
>Programing Languages and Compilers: 2-3 Months
not necessary, as long as you can use a language you can code
>Discrete Math and Probability (@ CS' level): 7-12 Days
>Algorithms, Automata, Computability Theory, and Complexity Theory: 1-3 Months
all uneccessary and you are not learning all of these things by self study while presumably doing school at the same time within 6 months.

>>12659212
>>12659365
Agree with second post, I'd just read up on the language that they were hired on, read up on basic data structures and algorithms, and maybe read the book "How to Solve It" so they go in guns blazing. Should all be doable in 6 months

>>12659212
He wants to be able to code, not get a mediocre, rushed look into a CS curriculum.
At most, he should use MIT OCW for data structures and a few classic programming books to learn.
>complexity
it's a wonderful topic but not necessary to program, especially if you're doing programming as a part of ME. If they had novel complexity constraints, it's literally easier to get a competent CS grad and bring them up to speed with the project (this is more common than you'd think, as long as some PE signs off that everything's working fine)

>>12645130
ok boomer

>>12647550
Just had an exam on this very book. Pretty interesting topic, I have to get around extending my lambda interpreter with all the cool types in that

>>12648333
have you ever solved a complexity problem before (theres a couple of problems fixed a year)? What makes you think you can reel in the big fish?

>>12659604
It's a good book, though I think Engineering a Compiler is probably the better text at this point.

How do people feel about the tiger book compared to Engineering a Compiler by Torczon

>>12659691
What's funny is how people think they can solve these big fish complexity theory problems from the simplest combinatorial principles as well..even though we're hitting a clear point where the enumerative combinatorics of the 70s can no longer adequately answer or formulate all our questions in complexity. Hell, that's the whole conceit with Mulmuley's GCT - the problem P vs. NP takes only elementary ideas to state, but highly non elementary methods in algebraic geometry, noncommutative algebras, cohomology to understand current work towards actually solving it

>>12659365
>>12659382
>CS is not code monkeying therefore we belong in /sci/
>woah woah woah, you can't actually expect us to learn theory, we're just here to code monkey and pretend to know math

>>12659935
don't act stupid. ME anon came in here freaking out about needing to learn 4 years of CS in order to do his job so it makes sense to advise that no they don't actually need to learn 4 years of CS and hes freaking out for no reason. if he wants he can continue learning CS but since its unlikely that thats what he actually wants to do then it doesn't make sense to tell him to cram 4 years of CS into 6 months of self study.

>>12657748
>ME anon
>>12659458
>ME
What makes you think he's in MechEng?

>it's literally easier to get a competent CS grad and bring them up to speed with the project
More like it's usually easier to find some other STEM major and bring them up to speed in coding rather than hope you can find a non-mathphobic cs major.

>>12659935
He was already hired for a junior position as an ME, and has 6 months to prepare. How much theory do you think he'll need? Here's your (you)

>>12659953
i was pretty sure he said he was ME in his deleted posts but i have since refreshed the page so I can't see it

>>12659968
No one quoted >>12657581 so I didn't see it.

>>12659935
Needing to learn how to code to get some mechE tasks done isn't the same as needing computer science lmao.
If he was needed to devise a novel numerical analysis algorithm for his work, it'd be a different story.

>>12659953
>bring them up to speed in coding rather than hope you can find a non-mathphobic cs major.
A lot of CS "majors" (the way you speak makes me think you're still in undergrad where it seems like such titles actually matter) aren't good at math, as most engineering majors aren't good at math despite ostensibly being more "knowledgeable," ie having another brainlet calculus course. The most mathphobic I've seen around are both CS and traditional engineering majors.

Stop being obtuse. There are clearly mathematically inclined people in every field, and engineering, both in undergrad and in the field, is not nearly as "mathematical" as you seem to be suggesting. Just because you struggled with Cauchy stress tensors or fluids doesn't mean nobody else can pick them up as needed.

Can anyone help me out here? Learning python. Isn't this supposed to be a dictionary?

>>12660293
https://www.programiz.com/python-programming/methods/built-in/zip
?

>>12660293
doesn't the zip function do that

>>12657748
>>12659212
>>12659365
>>12659952
Thanks for the advice anons
Sorry for deleting the post, wanted to post again with more details but I slept

>>12644893
What are the best books on information theory for layman? I love woo-woo, so give me the woo-woo ones too.

>>12660779
https://www.amazon.com/Mathematical-Theory-Communication-Claude-Shannon/dp/0252725484

>>12660293
It's zip()
https://www.w3schools.com/python/ref_func_zip.asp

>>12660779
https://cstheory.stackexchange.com/questions/37123/what-are-some-standard-books-papers-on-information-theory

Please comment on my presentation

lain chan /λ/res/9558.html #19751

>pic related: my automaton outputting a cotangent graph from no trigonometry input

>>12654284
Play with a radio some.

>>12659382
We all know you're wearing programming socks and belong in >>>/g/

>>12648881
What is the point of getting masters? If you have an IT job, it shouldn't be too difficult to transition to software job. If you want knowledge, self study is good enough considering you have an IT background. If you want research, start with knowledge then figure out a better approaches to current rule of thumb.

For IT can't you just get certs? What's the point of an IT degree?

>>12661269
needing to learn to program for an ME job isn't the same thing as needing computer science...

>>12644893
>Complete self taught pathway: https://teachyourselfcs.com/
Started the nand2tetris book from this link. The exercises are very satisfying.

>realized my stock options at my company are valued at 200k, and still pre IPO
I might become a millionaire and achieve my dream of going back to school and doing CS research

>>12644893
Stealing this pasta for purposes.

>>12662745
why do you need to be a millionaire to do research? just so you can live comfy for a while while doing research?

Should I change my major to cs from psychology even if I know fuck all about computers. I was interested in becoming a therapist but the pay is kind of shit.

>>12663402
Being self funded is a good feeling when you're doing research or going back to school. Otherwise, you're competing to get extra money / stipend / fellowships

>>12662745
Good luck. I wanted to do that but company failed to generate decent turnover so shit hit the fan

>>12663402
wife, kids, mortgage, want to eventually own lots of land. I don't really need much other than my computer & office, but I have responsibilities.

>>12663603
thanks annone, don't forget, we're all going to make it.

Too many books to read

ic

>>12645130
don't read this pile of parsing, there are better resources for compilers

>>12665314
stop reading
start making

>>12667473
do both

>>12644893
What should a CS degree look like?

>>12648545
obviously just double major

>>12669273
sometimes i wish i wasn't a lazy faggot and double majored. halfway through third year I'm realizing that I want to go into theoretical stuff but my CS degree has very little math requirement, and now I am too late and too busy to double major.

>>12669436
How are you too late? You would have had to take the classes anyways, just take them later

>>12669443
thats a lot of courses. I'm not aiming to be a super senior.

>>12669478
also i have a shit ton of debt, i need to pay for less classes and get out as soon as I can and hopefully return for grad school later.

heh

>>12644893

lol fuk this