/sci/ - Science & Math

thanks for making this
>any question
Python 3
why won't my shit run right? I've tweaked everything I can find but just make new problems. It just repeats the same randomly generated float infinitely

import random
import math

#x= random.randint(1,100)
x=100.0
print(x)

g_old = random.randint(0,100)
g_new = 0

while abs(g_new *g_new - x)>.01:
g_new = (g_new) /2,
print (g_new)
print('square root found, program finished)')

>>10523880
never mind, I guess rubber ducky technique does work,, this did the jo>>10523880

import random
import math

x=100.0
print(x)

g_new = random.randint(0,100)


while abs(g_new *g_new - x)>1:
g_new = (g_new+x/g_new) /2
print (g_new)

print('square root found, program finished)')

>>10523880
g_new is 0, and being divided by two, staying 0

Even if you use g_new = 100, dividing by two won't work. It will take this route
100
50
25
12.5
6.25
...

Basically, it will skip 10 entirely

I'm assuming what you tried was to do a binary search for the square root. Here's some pseudo code to represent how it actually works

[code]
function midpoint(a, b) {
//code
}

function sqrt(n) {
left = 0
right = n
approximation = midpoint(left, right)

while approximation is not good enough {
if approximation is less than actual {
left = approximation
} else {
right = approximation
}

approximation = midpoint (left, right)
}

return approximation
}

>>10523980
yeah I messed up a bunch of things. Thanks. babbys first python

>>10523980
>>10523952
Let's combine all the knowledge:
https://ideone.com/V45hbM

Is possible to use OpenGL commands with Pygame?

>>10523860
i wish this board had more serious programming discussion, the technology board is disgustingly anti intellectual

>>>/g/
what the fuck is wrong with you people

Is Project Euler the best place to start?

>>10525202
fuck /g/.
There are only corporate shills, trannies and trolls. In that order.

/sci/ is the new CS board. It's the only science that is relevant any more anyway.

>>10525202
forgot about CIA niggers.

>>10525214
If you like computation, yes.
But I like to make 'practical' shit more when I learn a language.

CS is not science
leave, >>>/g/

>>10525238
Like what?

>>10525244
Basically everything I use daily, file manager, photo viewer, web scapping, auto backup, etc.
I also use this book https://www.amazon.com/Exercises-Programmers-Challenges-Develop-Coding/dp/1680501224 when I run out of idea.

>>10525601
Thanks

>>10525224
what's the deal with all the trannies on there anyway, i don't remember it being like that a few years back

>>10525214
Depends. If you like hard math, I think you'll enjoy many later project Euler problems. I started with pe, but eventually transitioned into contest programming. I still wish I started using websites like codeforces, usaco, and topcoder earlier. If you live in the USA, start with usaco. Otherwise, I prefer code forces. I personally am not a fan of the open ended process in PE, where a single answer to check your solution can lead to hacky and suboptimal solutions. However, with real contests that's not the case, as they will check your solution against many test cases. Being good at contest programming is a very useful skill to have for colleges and jobs.

>>10525214
For general problem solving
https://open.kattis.com/
For interview prep
https://www.hackerrank.com

where can i learn arduino coding, this shit is so hermetic.

what's the future for graphics programming?

I want to master in it.

>>10526101
Graphics or visualization? Vulkan for the former, some JS framework for the latter.

>>10526106
graphics. I know visualization has a lot of potential in the future but I'm worried about graphics.

also http://www.anyonsys.com these guys look cool and they are looking for a vulkan dev in my location. I wish I was gud enough to apply

>1976 Matrix Singular Value Decomposition Film
https://www.youtube.com/watch?v=R9UoFyqJca8

>>10526096
>arduino coding
>hermetic
oh my fucking god

>Billionaire Mathematician - Numberphile
https://www.youtube.com/watch?v=gjVDqfUhXOY

>James Harris Simons has been described as "the world's smartest billionaire", amassing a fortune through the clever use of mathematics and computers. He is now a renowned philanthropist.

>>10526096
>arduino
>hermetic
if post==bait then kek=11
else this:>>10526738

>>10525102
https://www.willmcgugan.com/blog/tech/post/opengl-sample-code-for-pygame/

>>10525920
As someone who interviewed at Google.. the problem is it isn't a meme. It's a genuine element of fitting in in a FAANG, and certain communities are really big on that as a goal.

>>10526946
What's a faang?

>>10523860

>>>/g/
>>>/out/

I have two 5000x5000 matrices A and B that contain values up to 10^9+7. I'd like to compute A*B mod 10^9+7 but any software that I use is either too slow (integer python libraries) or too inaccurate (matlab, numpy). Does someone have an idea what else I could use?

>>>/g/

>>10526972
A C program you write yourself

>>10526972
Either C or Fortran (easier). I would see if PDL can interface C and Fortran for that though.

>>10526101
>>10526111
bumping my question with cute

>>10527150

gtfo

>>10525224
but at least /g/ lets you format your code correctly

>>10527163
what the fuck is wrong with you?

>>10526972
if you use a modular multiplication algorithm designed not to overflow you can probably do this in whatever language. 250 billion multiplications won't happen quickly but should finish in a reasonable amount of time. If you need to do this repeatedly or really need it to be as fast as possible I would look into GPU computing (CUDA, OpenCL)

>>10527278
>>10526972
I'd suggest vulkan instead if you just want to compute one simple thing extremely fast. OpenCL and CUDA have a lot of overhead compared to vulkan.

>>10527865
>vulkan
isn't that for 3d things? I'm a new random anon btw

>>10527918
It's a general API for not just graphics cards, but all general purpose gpus, the spec doesn't even require them to have the capability to draw anything to a screen. So consumer GPUs, mobile GPUs, enterprise GPUs without any display outputs for machine learning or whatever the latest meme is, it's a low level API for all of those, just barely above the mostly proprietary drivers of those GPUs.

It allows you to do manual memory management on the device, write to processing queues from multiple threads, and use compute shaders for general purpose programs that need a lot of brute force.

>>10527150
I work in AR. It's pretty cool.

https://youtu.be/LblxKvbfEoo

It's mostly embedded systems work (units API's, C++, Yocto and deeper linux/sensors/logic/fpga programming) but it's a good niche.
I work in the Bayesian inference end, though, not graphics.
https://en.wikipedia.org/wiki/Simultaneous_localization_and_mapping

>>10528059
That sounds very cool. Do you think AR is viable or will be viable in the future? From an outsider's perspective it looks like a entertaining meme but surely it's more than that.

>>10528204
It will surely come, although I don't know how prevalent.

Your mom will definitely want to check out if her Ikea furniture fits her living room before she goes and buys it, and stuff like that.
You can also trivially do stuff like facial recognition and query facebook and have peoples names and other information hover over peoples heads. Or tell you how fast everything is moving (cars on the street, say).

Or, if you wear the AR goggles, do volumetric reconstruction of the places you walked by.
https://youtu.be/c2wgrF08d1g

>>10528239
Nice. Well, you gave me an interest for AR, thanks AR anon.

>>10525920
dude just try to post anything remotely making fun of trannies or sjws. you will get instabanned.
happened to me twice already. It's why I left that shitty board.
I was only joking anyway. and its fucking 4chan ffs.

>>10528239
>volumetric reconstruction of the places you walked by.
that's cool, like uncovering fog of war in an RTS

>>10523860
/g/tfo

I want to get into c++, c# or Java but have literally no idea of any of these languages. Any good books suggestions for beginners?

>>10528492
The C Programming Language (Second Edition)

>>10528492
They are extensions on C.

Just watched the Coding the Matrix lectures.
What should I study next to continue learning about linear algebra for computation?
interested in basically all the things

>>10528586
>Coding the Matrix
redpill me on this, anon plz

>>10528613
oh it's the best. applied linear algebra with a heavy focus on applications and implementation

lectures https://cs.brown.edu/video/channels/coding-matrix-fall-2014/
class site with links to slide decks https://codingthematrix.com/
course textbook https://www.amazon.com/Coding-Matrix-Algebra-Applications-Computer/dp/0615880991

>>10528586
Deep Learning is mostly about deep systems of linear equations. Stanford has a good class videos for both CV and NLP.

>>10526972
You could try Julia. I'm surprised that numpy would give you inaccurate results for integers though, are you sure you are using the right data type?
If your data is actually floats, then you might be running into numerical stability issues.

>>10528621
thanks for that!! Looks very useful

>>10528492
C++ Primer Plus (5th Edition)
https://github.com/henear/Cpp_primer_Exercise/blob/master/C%2B%2B%20Primer%20Plus%20(5th%20Edition).pdf

C# 3.0 - The Complete Reference - Herbert Schildt
http://zeus.nyf.hu/~bajalinov/my_special/SW/C%23%203.0%20-%20The%20Complete%20Reference%20-%20Herbert%20Schildt.pdf

Introduction to Programming Using Java 8.0
http://math.hws.edu/eck/cs124/downloads/javanotes8.pdf

>>10525920
Transsexuality is the new trendy way of describing mental illness. It's the 'catch all' psychiatric solution to everything. It's honestly on the rise, in sweden it's increased by like 1000% or something among kids (from maybe 20, to 500, so it sounds worse than it is)

>>10526106
Why JS? It's slow as hell. Web assembly maybe, in that case (but i suppose the framework would be built on wasm)

>>10523860
Given a linked list, how would you add a new node to the front of said list.
In c++

>>10529270
Append the previous front node as the child of the new front node?

>>10529270
allocate new node whos next points to current head and then define list as that new node

>>10529270
newNode.next = startNode;
startNode = newNode;

Will SICP help me be a better programmer than CS monkeys?

>badoo
>someone sends a message to me
>i click on their chat box
>my browser sends a POST request to the server
>after it I can see the message and the person now know I've seen because when I received the POST response, the server detected I saw the chat and marked it as read
my question is: how to see the message without the server knowing I did so? Can I receive the POST response without server knowing it?

>>10529564
probably not, but it's still a good exercise

>>10529576
Possible if there is a security hole on that server, and you are a hacker.

>>10529576
depends if the data is pre-loaded in your browser

>>>/g/
fucking leave, faggots

How do I make a probability density estimator using neural nets?

>>10529547
> dot .
No! NO!!!

>>10530276
You make a label a fuckton of data

>>10523860
I'm making a program that can dove Newtonian physics problems. Cool huh?

>>10526101
Ray tracing.

>>10526972
The BLAS library has the fastest implementation of matrix multiplication

>>10529564
SICP is a foundation for becoming a better computer scientist, it's an entry-level text. although I believe HtDP does a better job. both good books though.

>>10529269
Because viruses?
HTML5 is the way.

>>10530276
Why would you want to do something with neural nets that doesn't require neural nets?

>>10523860
Isn't programming nowadays more a craft than art?

You don't need calculus, algorithms, graph theory etc. You just need practise and knowledge of your tools (hello shitton of JS frameworks).

Why is it such a well-paid job?

>>10530491
>Why is it such a well paid job
It sure as hell isn't. It's a bloated field that has too many wannabe professionals in it.

>>10530491
Because it's often hard and soul crushing. Hard in the sense that you are expected to be capable a huge range of topics and technologies, soul crushing in that a majority of that knowledge becomes invalidated within a decade. It would be like learning a new notation for Physics every few years because some suit thought this new one looks better on a whiteboard.

In reality, most programmers are not professionals. They aren't given the time to master all of their tools or the mentorship to learn proper technique and discipline from an expert. Instead you have an entire industry of novices routinely repeating the same mistakes over and over because the needs of industry outweigh all other factors.

>>10530491
programming is a tool what you do with said tool is what matters. This is why you need calculus, linear algebra, etc.

Is it stupid to write a doubly linked list where four lists are attached to a sort of "hub" in the center? Like a wheel with spokes (and the spokes are doubly linked lists, connected to the central hub). Wouldn't this be more efficient ?

>>10531435
It’s the simplest way to represent a graph. Whether it’s the best depends on the application, how much memory do you have, what assumptions can you make about the data. Without those, a linked structure is usually the best choice.

>>10523860
I’m planning to major in computer science. I’m pretty nervous whether I’m too brainlet for it. How fucked am I if I work hard and try to study for hours everyday?

>>10531778
you might not enjoy CS if you think it's just programming.

>>10531785
It’s the math I’m scared of. I’m willing to put in work to learn it but I’m worried about getting btfo. I’m trying to learn some Calculus 1 from Khan Academy now and over the summer to get a head start but it is a struggle to wrap my head around it sometimes, like reading another language. And people here say Calculus is the easy part.

>>10530463
Should I do both then?

>>10531804
Math is a notorius subject of being really confusing until its not, then what you learned is the easiest thing in the world. You're not behind or anything, just keep at it and youll get it.

>>10530491
Code monkeys aren't that well payed. 5g developers are.

>most cs programs in the united states don't have multivariable calculus or an advanced course on linear algebra

these US universities are the reason why many people see CS as a joke.

What should I call a function that takes in a collection of sets and two optional parameters N and M, and returns a set of all elements that appear in each of the input sets between N and M times?

Currently I just call it "intersection" (and describe it as an "arbitrary intersection" in a comment), but strictly speaking it's only an (arbitrary) intersection when N == M == length(collection_of_sets), which is the default. On the other extreme when N == 1 and M == length(collection_of_sets) the result is an (arbitrary) union. I'm not really sure what the call the cases for other values of N and M / the general operation represented by the function.

>>10532617
no way

>>10532666
How are you doing it?
How are the sets implemented?

>>10532676
The implementation is through bitmap indexing but that shouldn't really leak out to the function name.

As I've said, N and M are optional (and equal to the length of the input collection) by default, which means the result is essentially a set theoretical (arbitrary) intersection. But due to sampling/sensitivity errors I'm usually interested in elements that appear in say 95% of sets rather than in 100% sets like in a proper intersection.

>>10532690
I'd call it countRange, decorate it with doc comments

>>10528204
Working on AR smart phone apps. The technology is already cool for basic applications and it's a fun thing to work on.
The real problem with AR that prevents people from doing anything really revolutionary though is accurate positioning.
All the cool Aftereffects promo videos you see on Youtube, of people running through a shop or a street and it's all showing you exactly where to go and pops up information right where it should, all that.
That's all a complete fantasy if you don't know where the user actually is within about a square meter and where he's facing.

GPS is absolute garbage for AR applications. Maybe the military one is better. But the public one is way too imprecise. And you can't use it indoors. Same goes for the compass in your phone. Stuff like beacons is even worse.

Apple has a feature called AR world mapping which does basically this out of the box: >>10528239
And it's really impressive and it repositions everything where you set it up.
But this visual approach all starts breaking down when you take it out of the lab and into the real world where you have different weather, different light conditions, different times of day, lights off or on, stuff getting reflected, objects getting moved and in general people or objects moving in your way. And on top of stuff not getting detected, you'll have to worry about the wrong stuff getting detected as well.

Sure, you could bypass all this, by telling the user at one point: "Put the phone down in this frame that is perfectly aligned and press this button when you're ready." But that's just unsexy.
Even if you would then show fantastical dragons that guide you all the way through the store to your cornflakes, nobody would ever actually use it. People hate these sort of set-ups. Also they'd see pretty quickly that it's not a proper solution. Turn around and press the button and your corn flakes are in the parking lot.

Also you look like a dork looking through your phone all day.

>>10530490
Trying to catalog various methods and alternatives to estimating probability densities. I don't care if it spits out shit results i just need a way to pull it off.

>>10523860
so we all agree C is the best language right and you should do everything in it

>yfw u realize C is also the most environmentally friendly progarmming language as well
https://thenewstack.io/which-programming-languages-use-the-least-electricity/

feelsgood

>>10532672
it's true. They take up calc 1 and 2 (single variable calculus) with one course on linear algebra.

bump brehs

>>10526086
Recommend more websites like this.
Also, what's the To-go resource to learn algorithms?

>>10534333

>>10531435
>pic
I remember that thread.

>>10531754
>It’s the simplest way to represent a graph.
Fuck that. Representing a graph by linking nodes with pointers is fucking stupid. Store the edges and nodes separately.

>>10523860
Do you know anything about RSA encryption in JAVA....more particularly android?

>>10534581
So you have an edge object for every connected pair? How is that an improvement? Seems like it would take more memory with no benefit.
Unless you plan to store everything in flat arrays, but that adds complexity to the implementation.

Does it make me a lil bitch if I have to whiteboard out what I'm doing? I'm going through one of those online learn coding things and at first I could do them in my head and go straight to typing but now I can't, is this normal or abnormal for professional software engineers?

>>10534726
Whiteboarding is an essential interviewing skill.

>>10534665
Look up adjacency matrices and adjacency lists. Every time someone represents a graph with a bunch of cyclical pointer usage I die a little. I don't even care that it's more efficient. The clarity is what matters.

>>10534734
ok. is it just for interviewing tho? outside of interviews will it make me look like a mouthbreathing code monkey rather than a 100x god of programming

and when I say whiteboarding I mean planning things out on paper, not necessarily on a board

>>10534799
no, it's normal and professional to do so fot complex tasks. (if you work in a specialised CS field notably)

>>10534772
An adjacency matrix is inefficient in a sparse graph and scales terribly.
An adjacency list can still have the issue of cyclical pointers once you start following them through the graph.

How similar to math is programming?

>>10534847
In theory it's math. In practice it's often closer to plumbing.

>>10534862
I like math but hate plumbing will I enjoy programming?

>>10534863
you may enjoy the more theoretical parts of it.

>>10534865
Set theory and ordered bears?

>>10534799
People who code without think the problem through are worthless. Imagine being given a math problem and then immediate just writing numbers in the hope it gives the solution. Whiteboard any complex problem and think before you code.

>>10534842
adjacency matrices are better for weoghted graphs, and since its a matrix you can do linear algebraic operations on it, like find the transpose of the graph (all edges reversed) which is useful for things like finding strongly connected components

>>10523860
will i ever actually use discrete structures in life

>>10535117
No, but you will use them practically everyday on the job.

>when you solve your first problem on Project Euler

>>10527190
not worth it

>>10526946
>faang
>full of trannies
It makes sense. If you wanna work at one of these companies then obviously you have some issues.
Your main issue being that you are morally bankrupt and the second issues is that you care too much about other people's opinion.
It's just life's way of showing you what happens to people who care too much about others opinion and are morally bankrupt(Hint: they become trannies).
God works in mysterious ways indeed.
I hope you learned your lesson anon.

>>10523860
HOW DO YOU GENERATE AND STREAM AUDIO IN JAVASCRIPT???
I've looked at the standard library, but it doesn't have any easy way to implement the sorts of ring buffers I use in OpenAL.
THANKS IN ADVANCE!

>>10536532
in the browser?
look at web audio api on MDN.
there's an example there too. It's confusing, though

>>10529270
Node * newNode = new Node();
newNode->next = head;
head = newNode;

>>10536044
I know this feel deeply

>>10537844
Leetcodes are so much harder. Growing up sucks.

Oh I didn't know there was a programming general
(Are algorithms questions supposed to go here?)
>>10540482

>>10540500
Can I see code for this?

https://github.com/achael/eht-imaging
>python
disgusting
>GPLv3
patrician

The current FPGA class I'm attending is working with Matlab to show filter design. One of the functions used is "sfi"; however I'm using GNU octave; and looking for a Fixed-point toolbox; I stumbled upon the "Fixed" toolbox, but it looks like is no longer mantained, as it is not available in Octave Forge. I approached this thread to ask, if you knew of any other alternative.

>>10543233
Did they used a super-computer to process all those data?

Just passing by to agree that /prog/, or literal applied math with abstract words representing operations, is far more /sci/-related topic while "technology" as in actual physical technology is more /g/ related. Also it would be good to separate CS from the shill-containment shithole that /g/ is and have it come here instead as this board is dead anyway and anything that breathes some life back into it is welcome

>>10545429
No, just a few macbooks. Such is the power of Apple.

>>>/g/

Guys ive got 3 houses, each with fixed and variable loads. each with solar generation on the roof and 1 wind turbine, how would you schedule the variable loads?

>>10545525
to specify, i need to find at least 3 different approaches used in the decision making process (e.g., one deterministic hierarchical, one Monte Carlo simulation, one optimisation)

is there anything more soulless than spending your waking hours staring at a screen of syntax all day?

>>10536044
>>10537844
solved like 20 eulers until the math got too complex years ago, dont have the sources anymore but would it have any merit in CV?

>>10545541
>tfw programmer's block

how many fucking method names and functions do I have to memorize. I'm going through a tutorial and I keep needing another obscure one I learned 5 or 6 lessons ago. I keep thinking I should just put them on anki but if no one else actually memorizes these and just googles them when they need them I wouldn't bother

>>10545829
I guess it wouldn't hurt to make a portfolio of them and have it on your Github.

>>10547074
just practice enough until you remember them