Ruby is my wife editon
Stale bread:>>53995476
what ya workin/g/ on?
Making a neural network to classify points having their coordinates as input. My first attempt at neural network anything.
White background is predicted as red, black background is predicted as blue.
>making it before the bump limit
>by a tripfag no less
enjoy you're no replies
first to say wumbo
Feels good no one has made fun of my programming language Java in these threads in awhile I see you guys have come to your senses and realize Java for all it's glory
>>54000925
>2D
lmao
>>54001008
it's not the real OSGTP
>>54001043
If you have less than 512 dimensions in your input, don't bother.
>>54001054
who?
>>54001058
>don't bother
But how will I learn?
>>54001076
By starting with classifying mnist digits.
>>54001020
kill yourself.
>>54001086
One step at a time anon.
Are we going to just sit here and shit post all day long or are we all going to collab and program an application together to make good software?
>>54001130
mnist is step 0.
>>54000834
I hate PHP too, but please let this be fake.
> function hashing mechanism was strlen()
anyone tried using Crystal?
>>54001066
the burger with the !!Ju67JuF6jBe tripcode
>>54001151
You are in such a rush, anon. Walking slowly is preferred here.
>>54001148
start something and people might join in
>>54001175
No it's clearly the dude with the ##cocks tripcode.
>>54001187
Should I just make a git project and have a quarter finished project and see if people would be willing to contribute?
>>54001221
that's probably your best bet
>>54001233
I am curious would any of you guys actually contribute if one of us did that here?
>>54001148
>>54001187
>>54001221
>>54001233
/dpt/ in incapable of collaborating. Many people have tried before and failed miserably, because /dpt/ is just not suited to that sort of thing.
You would have a better time trying to herd cats.
def is_prime?(number)
2.upto(number-1) do |x|
return false if number % x == 0
end
end
def prime_range(start, finish)
start.upto(finish).select do |x|
is_prime?(x)
end
end
>>54001279
It's shit.
>>54001247
if it's something interesting and in a proper language
>>54001269
people made tox. it's shitty/overrated but it's something and that douche kid got a few thousand of donation money
>>54001304
Like what? What kind of projects would you guys contribute too?
>>54001304
>people made tox. it's shitty/overrated but it's something and that douche kid got a few thousand of donation money
The exception rather than the rule. Tox distanced themselves from /g/ VERY quickly and is hardly a /g/ project.
>in a proper language
C is the only viable language for /dpt/.
So, I'm trying to use the openhardwaremonitorlib.dll to monitor my system in C#.
Now, I know nothing about C# (I'm a C++ guy and only barely remember how to do OOP), but apparently it's either that or VB if I want to find any information at all on how to do this. And VB doesn't work because I want to do serial communication with an Arduino once this part of the project works. Apparently this can be done, but only with workarounds.
I already found a way to include the .dll in my references and found out how to view the structure of the .dll.
I found someone's small program and tried to run it, but it doesn't display anything for me. I also tried to edit it to align it with my style, but then I realized that I don't even know how the fuck C# software is structured.
The guy used a hardware.Update(); method call and this threw me for a loop, until I realized that it's not a function call for an Open Hardware Monitor object, but rather a variable that he declared in an obtuse way, (inside of a foreach loop - also a concept that I needed a few seconds to actually comprehend) with Update apparently being a system function that I'll now need to look up.
I seriously have no idea what the fuck I'm doing.
>>54001328
>C is the only viable language for /dpt/.
Idiots who say that will not participate in anything anyway.
>>54001342
I'd help if I could.
>>54001319
idk, something "lulzy", but yeah it's a stupid idea to begin with to try to force a collab when you don't have anything to collab on
>>54001247
I'll make the logo! :^)
>>54001342
Every other language is either too niche or hated by /dpt/.
Picking anything other than C would make you lost a massive amount of potential contributors.
>>54001319
>>54001367
Wouldn't something like a system monitor widget with built-in Dogecoin miner and instant messenger be right up /g/'s alley?
>>54001379
people don't use C or (rightfully) think it's shit. the people who unironically use C are shit at programming
it should be in C++, people can still write parts of it in C or C-style C++
>>54001367
Well here are my ideas since I am a great idea guy
a web browser that plays anime videos
a sound wave frequency detector
an internet tabs management system
>>54001395
Here is your (You)
>>54001395
>people don't use C or (rightfully) think it's shit
*people don't know C or (rightfully) think it's shit
>>54001397
Those ideas are pretty bad.
>>54001397
>a web browser that plays anime videos
Where can I sign up
>>54001410
Well lets hear ideas you guys would like to contribute to
>>54001397
It should be some kind of botnet for Raspberry Pis.
Everyone can then do their share by extending the botnet whenever they have cash left over.
Maybe a neural network for captcha solving.
>>54001423
I'm not an ideas guy, I'm a programmer.
>>54001379
>>54001395
>>54001342
>>54001328
Lol this is why /dpt/ can't do shit, every attempt at collaboration devolves into "muh language is better" before anyone even comes up with any ideas.
>>54001405
This. Just look at ANY of these threads. The ctards always know the least about C.
>>54001439
Then shut up and program my ideas.
>>54001441
Shouldn't we decide on language based on our needs, not decide on our language and then decide on our needs?
>>54001450
Just because I can't cook doesn't mean I can critique your food. Make me something good and I'll program it.
>>54001455
Whoever starts the idea has to choose all the tools and languages and write a first deliverable. At this point, anyone can come in and participate. The project leader (i.e. first contributor) gets to accept or reject PR's.
>>54001431
if we're gonna do something like a botnet, maybe a captcha solver, could we use ethereum somehow so the people who run botnets get paid for it
>>54001467
We need a great leader
>>54001447
What on earth are you talking about?
>>54001484
>being this delusional
>>54001484
Are you stupid or pretending?
>>54001503
>>54001507
I know C and I think it's a great language. Stop projecting so hard.
>>54001475
>/g/ makes a money-generating Captcha-killing botnet that can be run on thirty dollar machines
This sounds like the story of a future cartel or Mafia-like organization.
I like this.
Add something funny that creates enjoyment as a secondary function and people will eat it up.
Something that is obviously doing something. Like an optional visualization of the whole process or a counter for how many nodes there are and how powerful the system is and how high its success-percentage is.
PRODUCTS = [
{ name: "thinkpad x210", price: 220},
{ name: "thinkpad x220", price: 320},
{ name: "thinkpad x230", price: 500},
{ name: "macshit", price: 3000},
{ name: "macshit 2", price: 1100},
{ name: "dell latitude", price: 650},
{ name: "dell inspiron", price: 830}
]
query = {
price_min: 300,
price_max: 950,
name: "dell"
}
def search(query)
PRODUCTS.select do |x|
((query[:price_min]..query[:price_max]).include? x[:price]) &&
(x[:name].include? query[:name])
end
end
>>54001431
>>54001475
How would you guys feel about the captcha solver idea?
More ideas! Keep them coming
>>54001441
you'd need a shitload of training data to keep up with captcha. far more than any one person or small group could ever possibly manage. maybe there's some clever way to outsource the training of neural networks on a large scale
>>54001520
Confirmed for never having programmed anything beyond hello world in C. Protip, Timmy, hello world is not the peak of programming expertise.
>>54000925
Maybe I shouldn't give up on neural nets, although honestly idk what to do with them other than my failed attempt
It gets it right sometimes; it's probably just my shitting training set
>>54001565
oops meant to link this
>>54001431
>>54001542
Where would you get the necessary database?
>>54001547
If we do this, I vote for pic related as a theme. He was a prolific autospammer on /a/.
Everyone hated him until the mods banned loli. Then he spammed loli for us.
>>54001535
Everytime the captcha is solved a funny lol cat picture will appear on the user's screen people love internet cats this would work
>>54001571
I don't believe I've ever seen somebody project so hard.
Here is your (You) anyways.
>>54001565
There are open databases of captchas. In principle this is enough. The whole point of moving away from hardcoded experts systems toward learning systems is that they are fantastic at generalization.
>>54001565
that's literally what the captcha is
lol
>>54001565
4chan users could run it in the background and it could get training data in the same way that google gets it with each 4chan post. i could see 4chan users volunteering for this
>>54001565
But isn't Captcha our training data?
If it gets through, it was correct. If it doesn't, it wasn't.
It needs shittons of VPNs, though.
>>54001475
>>54001547
Let's make p2p neural net captcha solving botnet!
>>54001621
>each computer is a node
this is great
>>54001612
And never forget: 4chan is most likely one of the pages with the highest amount of Captchas solved.
It's also one of the pages with users most likely to support the destruction of captcha and enough autism to do double shifts for more captcha solving.
>>54001627
Can we have an addon for 4chanX (is that even the current one?) that taps into the network and watches people solve captchas?
That's in addition to the system itself learning by trying to solve captchas separately.
Not sure if this is the right place to post, but does anyone have any documentation for how to hook into the menu bar in OSX?
I want to be able to display a "Now Playing" field in my menu bar, but there's no widget to do so from mpd, so I was thinking that I could just pipe the information from mpc (the same way I would on linux) somehow, but I can't find relevant documentation to make what I want to make.
>>54001656
install linux
A better place to ask would be /desktop/ but only a few of them use OSX
It has been decided you all have a hard on for the captcha solver idea
>>54001580
What kind of network are you using? How did you train it?
I myself have absolutely no knowledge in subject. I saw how you build a network out of seemingly arbitrarily chosen network in some tutorial, and change parameters of nodes using backpropagation. The network in picture I posted is trained using those points on picture, fed into model in random order 80000 times.
My network has 2 nodes as input (coordinates), 8 hidden nodes, each multiples two inputs by a parameter and adds another parameter, the result of that is sent into tanh, and each of two output nodes adds 9 values - output of hidden nodes multiplied by a parameter and another parameter.
>>54001607
lol yeah that's the joke
>>54001605
my point is that captcha already trains networks, some of which likely goes into improving itself and making it harder to cheat with a weaker network. you'd have some catching up to do
>>54001612
that would be a good place to start
>>54001613
>If it gets through, it was correct. If it doesn't, it wasn't.
nitpicking, but that's not entirely true. captcha doesn't know you're correct in every case. it can only verify the cases it already knows the answer to, which isn't most of them. it has to have cases it doesn't know the answer to, otherwise it wouldn't be learning anything
>>54001397
>a sound wave frequency detector
You can use a Fourier transform and create a voice changer.
>>54001683
I wrote the alphabet into it 20 times as my training data
It's feed forward, back propagation of course, I'm to dumb for anything else
It ahs 256 input nodes, 450 hidden nodes, 5 output nodes (for binary 0-32, use 0-25 for alphabet), its linear -> tanh -> sigmoid, just used what trained quickly but gave me 0 < x < 1
>>54001683
Toppest kek. You're never going to get anywhere with less than 50 nodes per layer, even if you're trying to split a plane in halves; let alone if you're trying to do something remotely serious. A simple mnist sigmoid regression will use something like 1024 nodes in the hidden layer and 10 outputs, for a total of 814090 parameters. And that's the smallest real-world model you could possibly find.
>>54001717
I actually wrote it 40 times because right after finishing the first 20, I accidentally overwrote it and lost it all
took about an hour and a half to do it twice :'(
>tfw writing concise, clean and fast C code after months of OOP and C++
Is there any better feel?
>>54001664
I like OSX on my workstation for some of the productivity conveniences it provides over linux.
However, I don't want to use itunes since mpd is better, and I've already got a configuration that I like from my linux box.
The reason I posted here was because I assumed that finding resources for making a menubar widget for OSX would be more of a programming type question than a ricing type question.
>>54001035
>>54001058
>>54001086
>>54001151
Fuck off and kill yourself.
>>54001708
Found the inbred, everybody!
>>54001708
>you'd have some catching up to do
True, but can you beat the combined power of an expanding botnet coupled with autistic humans?
>>54001153
https://en.wikiquote.org/wiki/Rasmus_Lerdorf
Ok, here is the time line for the new project:This thread | Fantastical ideas. First set of logos proprosed.
Next thread | Github account set up; only contains README. More logos.
After that | Project spins off into its own thread. 100% logo shitposting.
2 days time | Project threads die off. No mention on /dpt/.
3 days time | Everyone has forgotten about it. Github only contains README and license.
One week | A single poster will ask what happened to the project; gets no replies.
>>54001725
Writing Python after using only C for two years was great
Not as good as that tho
OOP is the least pleasant thing to write
How would a captcha solver actually work to make one?
>>54000925
Good, anon. Continue.
>>54001722
that depends entirely on the problem. there are simple problems that can be solved to high accuracy with pretty damn small networks. it also depends on the application. you could prefer a smaller and less accurate network because it meets your requirements while having a lower runtime cost, for say a realtime application
>>54001745
Every single time.
>>54001755
it would take the captcha image as input
and output the text
:)
>>54001722
But my thing does split the plane properly..?
>>54001717
>5 output nodes (for binary 0-32
Can it really work like that? Would the quality improve if you just use 26 output nodes each denoting the probability of letter?
>>54001755
https://github.com/search?utf8=%E2%9C%93&q=captcha+solver
>>54001740
>Back when PHP had less than 100 functions and the function hashing mechanism was strlen(). In order to get a nice hash distribution of function names across the various function name lengths names were picked specifically to make them fit into a specific length bucket.
What the fuck, is this nigga serious?
>>54001772
I guess I'll try it. I read a paper that used 7 nodes for caps, lower case, punctuations, and numbers, and it worked fine; that was my inspiration.
I didn't think that would work because the binary representation of a letter's index in the alphabet has nothing to do with the letter structure at all; I'll change it up and post results. Thanks for the idea anon.
>>54001717
20 times is far too little.
>>54001717
I recommend learning a little bit more about these before trying to implement one. There's a lot of poor design choices in your post, least of which is outputting binary (it might be possible for this to work but nobody fucking does it).
http://cs231n.github.io/
https://www.youtube.com/playlist?list=PLkt2uSq6rBVctENoVBg1TpCC7OQi31AlC
Here's a course that will teach you everything you need to know about how to do this shit.
>>54001767
Runtime performance is virtually never an issue. Even 20+ layers of serious convnets do a forward pass in under 0.3 seconds, let alone models appropriate for realtime systems. The only problems where you can use very small networks are either not ML problems (for example, RL systems can work in real settings even with really few parameters), or not problems at all (again, splitting a plane - you can do that in a single iteration with the right algorithm and you gain precisely nothing by using a neural network). In particular, if your problem is that small, an svm is way more appropriate than a neural network. It will even work on significantly less input data than neural networks with more accurate results. The reason svm's aren't used at scale is that they don't actually scale, while NNs do.
>>54001717
You'd have significantly better results by using 25 outputs, each denoting the likelihood that the character has 'this' label.
>>54001802
>binary representation of a letter's index in the alphabet has nothing to do with the letter structure at all
Well, neither has the arbitrary number you assign to letter.
But it still feels to me that predicting that the letter is 9th letter of alphabet is easier that predicting that 3rd bit of letter number is 1.
>>54001810
Not him but thanks for those.
I implemented my thing in C++ to get a feeling for it.
But literally everyone is doing neural networks with python.
I guess in the end I'll have to yield and switch to python also for a more comfortable experience.
>>54001864
It's a fact that it's significantly easier. You can even do cool things like smooth over the prediction with an SVM at the top to improve your performance (although nowadays it's not popular to do that).
>>54001580
Go with 26 output nodes otherwise you're wasting a layer for binary conversion
You could reduce the required number of nodes by using convolution as the letters can be at different positions
Either go with way more training points or way less nodes (write with a pen on your touchpad or whatever to speed it up).
Tanh and sigmoid are the same thing
>>54001864
http://neuralnetworksanddeeplearning.com/chap1.html
Search for "You might wonder"
Explains very well why, even in your context, you should use 26 outputs
>>54001916
There's caffe for C++, but I don't recommend it anyway. Try torch for lua or theano for python. Stay away from cntk and tensorflow, their public releases are not usable yet (except for toying around, which is fine for you I guess).
>>54001926
>Explains very well why, even in your context, you should use 26 outputs
I assume you misquoted because i'm the person who initially suggested him to use 26 outputs instead of 5.
>>54001916
Can we stop with the neural network meme? It's fucking embarrassing.
>>54001968
What are you embarrassed about?
>>54001827
0.3 seconds is not fast, and convnets would ideally be usable in realtime applications, so consider that maybe runtime performance is sometimes an issue. and i disagree anyway. i'm working on an ml problem right now that uses relatively small networks because i need forward passes on multiple instances with plenty of time left over per frame for the other aspects of the application
I'm in the middle of making a meme language /dpt/. What features should I include? Reply to this post with ridiculous ideas.
>>54002016
Object oriented programming.
>>54002016
C++ templates.
>>54002016
Everything is an object
>>54002016
Native serial interface to expand the computer's processing power.
>>54002016
make fuctions execute in reverse
>>54002016
Homoiconity.
Case-Insensitive symbols.
I'm getting a syntax error but I'm not using the < symbol anywhere near this code, any ideas?
sqlite3.OperationalError: near "<": syntax errordef calculate_average(obejct):
c1.execute("SELECT * FROM dataLog WHERE object=%s"%object)
data = c1.fetchall()
for row in data:
print(row)
>>54002016
user-definable swizzle operators
>>54002016
Make it Arabic and right-centered.
Use Kanji for all default functions.
>>54001987
Consider reading before posting next time, maybe it'll make you look less retarded!
Then again,
>uses relatively small networks because i need forward passes on multiple instances
Shows you're literally using your anus as a mouthpiece figuratively.
>>54002016
Linear and dependent types with system-F<: inference.
>>54002016
store functions in a hash table, use strlen() for hashing and choose names for the standard library in a way that allows you a good distribution
>>54001725
smart frogposter
>>54001968
I hear lisp good for neural nets? I've just started learning it out of boredom, does anyone have any lisp AI resources?
>>54002051
I'm new to this, but could it be that "obejct" should be "object"?
>>54002089
>I hear lisp good for neural nets
You heard wrong
How come none of you work on projects to make money?
>>54002095
Sometimes I wonder if I'm retarded. Thanks man.
>>54002100
you got me
>>54002089
It's not. This is an ancient meme where lisp was good for AI (never for NN by the way). Python isn't good for NNs either but at least it has several good frameworks for it. Lisp has literally nothing.
>>54002100
>>54002123
samefag
>>54001968
This.
Kids just think they're geniuses for programming a "neural network", it's worse than the fizzbuzz enthusiasts.
>>54002060
somebody's butthurt. it sounds to me like your understanding of neural nets is relatively narrow. i'm not the one that looks retarded here. you could try outlining why specifically you think i'm wrong instead of resorting to childish behavior
>>54002016
Japanese-based syntax only. No alphabet allowed.
>>54002144
Haven't laughed this hard all month. Thanks bro!
>>54002101
But I do.
>>54002155
>>54002136
>>54002130
>>54002123
>>54002119
>>54002112
>>54002101
>>54002100
>>54002097
>>54002095
>>54002079
>>54002155
samefag
>>54002141
>>54001968
This. The worst part is when kids have literally no clue but try to pretend they're experts, though, it's even more annoying than the retards who think they're geniuses for making toy neural nets.
>>54002159
not up to the task? what a surprise
>>54002175
>>54002136
actually no
>>54002195
ahahaHAHA autismatisimo
>>54002198
> >>54002100
> >>54002100 (You)
are you even trying anon
>>54002141
>>54001968
>>54002177
the worst is that troll that just post slides off of research papers and says it's better than ordinary programming
ML with infinite resources is like having an endless supply of pajeets and time. it's suitable for some things but for a lot of things it's still shit.
>>54002213
Don't respond, he'll keep shitposting if you do.
>>54002231
Don't fucking reply to me ever again unless you're contributing to the thread.
>>54002231
im actually reverse trolling him
>>54002227
ML completely wrecks everything whenever someone so much as tries to apply it to a domain, though. The only thing it's not the best at yet is language comprehension. In every other domain it's significantly superior to everything else, and in some domains, like vision and automated agents, it's significantly better than humans.
>>54002254
i don't think you understand how deep this hole goes
>>54002257
it's only good in very specific domains you fucking retarded memer. get ML to make a game, a website, an image editor, a movie, a dildo factory. you won't get anything good out of it.
>>54001803
I know. My hand was dying tho
>>54002282
It's OK pajeet, don't worry, I'm sure you'll still have a job in 10 years! Shh, baby, shh!
>>54002101
Because I'm a programmer, not a marketer. Let someone else figure out what people want to buy.
>>54002293
kill yourself retard
>>54002299
(You)
>>54002300
go back to >>>/n/ you piece of shit we don't need cancer on this board
>>54002325
>>54002322
>>54002309
>>54002300
>>54002299
>>54002298
>>54002293 (You)
sry wrong post
>>54002322
you're the cancer fucking retard. i acknowledge that ML is good in some domains but it most definitely is not good in "all other" domains like you're claiming.
>>54002339
Don't bother. Just ignore the memer kids.
>>54002346
>everything I don't like is a meme
>>54002351
if you think ML is good for everything you don't understand ML
>>54002346
>hating memes
get the fuck off my board normalfag scum
>>54002375
(You)
>>54002361
>>54002384
please edit this page to say it's good for everything
https://en.wikipedia.org/wiki/Machine_learning#Applications
Daily reminder this kind of cancer posting only happens when Yuropoors are at home while Americans are working or at school
>>54002403
why does his eyes look like that? i get the pale skin but is he on stims or something?
>>54000834
And Go is your wifes son?
>>54002416
europe made america from fucking scratch
>>54002416
You forgot one continent.
It's obviously Straya.
>>54002406
Have another (You), on the house.
>>54002449
literally delusional
>>54002434
The pupil always bests the master.
>>54002468
HIGH
T
E
S
T
>>54002468
>>54002455
>>54002452
>>54002449
>>54002437
>>54002436
>>54002434
>>54002431
>>54002425
would you please stop shitposting and get back to the topic this toxic shit has to stop
>>54002468
>>>/gtfo/
thanks, [s4s] board culture enrichment program
>>54002403
I don't want to smell his mouth
>>54001209
This guy?
>>54001952
Yes, was meant for >>54001802
Excuse the (You)
>>54002484
Fuck off cuck
What in god's name is going on in here?
I'm setting up a C++ testing framework for my company. I'd like to use IPC on the testing process, but I want it to be cross platform. Seems like Google has a decent IPC library, is there anything better?
>>54002693
>>54002130
Do you think common lisp is worth learning in 2016?
don't mind me, I'm just making this thread better with browser #112.
What's the best live c++ parser for sublime text 3?
there are too many to chose from
>>54000834
writing a web server in go
>>54001334
Why not use C++ to open the .dll with LoadLibrary and fetch functions with GetProcAddress?
>>54001745
sick, let's get started
>>54002806
Do you have no eye for visual appeal? It looks pretty ugly; I've always thought that when I've seen these. All it would take is swapping soe images and it would look much better. Except for the thread view, which looks awful.
"thats a long and boring story you dont wanna know"
so meta
>>54002912
>soe images
wut
I do agree that the static screenshots generally look like a clusterfuck, but when you're actually using it, things make more sense. in any case, it is somewhat user adjustable.
but the whole point is to cram as much shit as possible into a small area. If I wanted something spread out, with little in the way of info being displayed, I'd use the board's built-in catalog.
>>54002888
Because I literally don't know anything about .dlls.
Thanks for the idea, anon.
DB-related question here, graphs vs dbrms >>54000925
Linked it here because everyone knows this is where the competent fucks hide.
>>54003002
some images*, my bad
I meant something a little more modern; think about the bump, street cone, etc. And maybe replace the blue with a flat, solid color or something.
Just my opinion though, if you and your users like it then by all means stick with it
>>54001745
Is there ever a successful project from 4chan?
>>54003057
Are you sure you linked the right post?
>>54003078
>street cone
I mean't the red triangles with the exclamation mark, for bump limit I presume? No idea why I said street cone lol
>>54003078
those are the overlay images, and they CAN be toggled. I'd love to make each of their x,y,w,h adjustable and maybe even use custom images, but that requires more UI work and getting it to fit in nicely with the other options. I'm trying to keep the Options window small, with everything visible at once.
>>54003093
>exclamation mark
that's for threads on the last page. it means "this thread is about to die/fall off the board"
complete newb here
Is it better to use nested ifs or switch inside an if statement? How can I improve this code?#include<iostream>
#include<stdlib.h>
#include<time.h>
using namespace std;
class Rps{
private:
string userChoice;
int computerChoice;
public:
void game(){
//Storing userchoice
cout << "Rock ,paper or Scissor(all in lowercase): ";
cin >> userChoice;
//Generating computer selection
srand(time(NULL));
computerChoice = rand() % 10 + 1;
if(computerChoice <= 3){
cout << "The computer selects Rock\n";
}else if(computerChoice <= 6){
cout << "The computer selects Paper\n";
}else{
cout << "The computer selects Scissor";
}
//Outputing results
//UserChoide: Rock
if(userChoice == "rock"){
if(computerChoice <= 3){
cout << "It's a Tie";
}else if(computerChoice <= 6){
cout << "Computer Wins";
}else{
cout << "You win";
}
}
//UserChoice: paper
if(userChoice == "paper"){
if(computerChoice <= 3){
cout << "You win";
}else if(computerChoice <= 6){
cout << "It's a Tie";
}else{
cout << "Computer Wins";
}
}
//UserChoice: Scissor
if(userChoice == "scissor"){
if(computerChoice <= 3){
cout << "Computer wins";
}else if(computerChoice <= 6){
cout << "You win";
}else{
cout << "It's a tie";
}
}
}
};
int main(){
Rps game1;
game1.game();
}
[\code]
>>54003078
maybe this is more to your liking
>>54003091
>>54003057
meant to link >>54003028
>>54003252
That image example is easily represented in a relational database.
is there a way to use sed between two defined character positions, for example the second and the 4th character. i know i can do it UNTIL specific characters if i know what character i want, but i cant figure out how to do it as a range of unknown characters
I'm trying to do something small - display an image in dos in 320x200 256 colors mode.
Now I'm wondering how to convert a png to that color pallete, in order to get the color number of particular pixels. Then I'd simply hardcode it in any transfer it to video segment byte by byte.
Any tips?
>>54000834
>what ya workin/g/ on?+ 0 post omitted.
Attepting to make RoR work.
It doesn't.
>So lazy they haven't fixed a json-install issue on windows since 2 fucking years ago
when did you first realize java was comfy? is there a comfier language?
>>54003730
C++
>>54003272
It can be done, just like that, yes.
But consider that the "owns" relations or the nodes need to be generic. Still doable, but perhaps a graph db is better suited?
>>54003810
No.
>>54003730
Everything is comfier than jaba the hutt.
I gave up writing this script last year, but thought someone could maybe use this script / idea to create something better or at least working.
It is supposed to create a encrypted rar archive of a folder, split it into pieces of 50gb and then upload it to mega using multiple accounts.
https://github.com/FELIXakaFX/MegaBackup/blob/master/megaUP.sh
>>54003810
Graph DBs might be better suited to your particular application, if you're mostly looking at immediate associations amongst nodes and the nodes themselves are generic.
neo4j looks pretty cool, but I don't have much experience with graph databases. I'm unsure of their downfalls, or where they might be less effective. I'm curious about the actual physical storage and the performance therein, especially comparisons between SQL-based indexes (clustered) / SQL Plans/Statistics and however a graph DB would handle large-scale aggregations.
>Anna
>Frank
>Anne Frank
oy vey
>>54003730
Java is okay, but C# is comfier.
>>54004036
What is the .update() method supposed to do when used on a variable?
can someone help that's knowledgeable about linux shit
sed 's/PI.*Value://'
this removes everything between Pi and Value. instead of Pi i want it to be the first character, so i figured it would be /1g, but that doesnt work. it says unknown option to s
>>54004130
Assuming non-retarded programming, whatever the most logical assumption that can be made about what it is to 'update' that particular object.
These things are easily determined by navigating to definition of the object to beUpdate()'d
>>54003810table nodes:
id
--
0
1
2
table persons:
id name born twitter
-------------------------
0 dan ... ...
1 ann ...
table cars:
id brand model
-------------------
3 volvo v70
table relations:
id_src id_dst relation width
----------------------------------------
0 2 drives 1
0 1 lives with 1
0 1 loves 1
1 0 loves 0.5
1 2 owns 0.5
pretty generic
>>54004211
uh oh here's a correction:table cars:
id brand model
-------------------
2 volvo v70
>>54004211
>not abstracting the types of relationship types into its own table to avoid all that bloat
>Make a program to fetch conversations online
>Decide to make a simple fucking graph with edges (source -> target)
>Have to use a literal who program called gephi
>It's fucking java based
>Have to install fucking java
>After much tinkering around I manage to load the info into the graph
>Extremely hard to make a format that looks good
>Eventually it looks ok
>Try to export
>Program just hangs there
>Java consuming 90% CPU
>Even if I exported to json/gext/svg the javascript library to display that shit would just freeze because the sheer amount of data
All I wanted to do was a pretty graph that grouped people to who their talk with the most
fuck my life
/blog
I was reading about memoization and continuations. Rate my code for the Collatz conjecture problem.let memoCollatz =
let cache = System.Collections.Generic.Dictionary<_, _>()
cache.[1L] <- [ 1L ]
let memoize n r =
cache.[n] <- r
r
let rec collatz =
(fun n ->
let rec loop continuation n =
match cache.TryGetValue n with
| (true, k) -> continuation k
| (false, _) ->
let nextValue =
if n % 2L = 0L then n / 2L
else n * 3L + 1L
loop (fun accumulator -> continuation (memoize n (n :: accumulator))) nextValue
loop id n)
collatz
It saves the entire sequence and as it uses a linked list, sequences that repeat share the same memory.
>>54004232
Of course you can do that.
I only showed how to represent a graph with nodes in a relational db.
Today, instead of writing a base class and inheriting 2-3 classes from it, I chose to use composition.
Just a few structs inside another. Constructors and destructors still work and do their job. Combined with C++11's in-class (or in-struct in my case) initialization of members, I don't even always need constructors anymore.
It feels liberating, in a way. No fiddling with unnecessary OOP, dynamic_casts, no unnecessary pointers (smart or raw). Just... data. Straightforward. Why haven't I done this earlier?
>>54004265
>Decide to make a simple fucking graph with edges (source -> target)
>Have to use a literal who program called gephi
why does that follow?
Is it really true that openmp on windows doesn't support thread-local variables?
>>54004382
what?
>>54004441
why do you have to use another program if you want to make a graph? why not just make it yourself? or are you talking about a visual graph
>>54004398
>In OpenMP, one can also create thread-local variables by specifying them in a private clause on the parallel pragma.
>>54004489
i mean a graph like the pic i posted
rate my bomb
>>54004515
i thought that was just a piece of abstract art
if i wrote a program in python and then translated it into C, without fancy optimization techniques, just normal reasonable code
will there be a noticable speedup in the average case? speak from experience please
>>54004603
if you translate it
>>54004603
i've spent years translating python to C and can tell you that yes, the code has a noticeable speedup
>>54004603
>in the average case?
Noticeable? Nope.
>>54004543
Noice bomb Ahmed, what micro are you using?
>>54004603
Written by hand or using Nuitka or similar? if first, yeah, it will run faster.
I had to teach myself more or less everything I know about languages, what are some things I can read to understand terminology and shit better.
Namely, I'd like to understand anything that >>54001827 said
>>54004640
I mean the average case for programs that have to deal with some obscenely large input and a lot of going through loops and computations etc
>>54004543
shitty jumper wires and no current limiting resistors out of 10
use 8 shift registers to create 8x8 led matrix that doesn't have to be demultiplexed, then send frames from pc using usart or vusb in order to create animations
>>54002016
I remember a bunch of retarded ideas I had in hs.
>Every function had to be drawn out as a rectangle. The compiler would push the functions together like puzzle pieces and fail if it wouldn't fit in perfectly.
>A language that would behave like a dog, having to make the same call over and over again (a random number of times, mind) until it finally retrieved. You would also have to keep reminding it that it was good code or it'd get sad.
>Compilers interpreting sheet music (but then I found out someone actually did this)
It's a good thing I wouldn't have know how to even go about making it if I had actually tried
>>54004641
>what micro are you using?
attiny4313 currently, but just ordered some atmega8a which I will use the next weeks, because of uni.
>>54004731
>shitty jumper wires
hail china :^)
>and no current limiting resistors out of 10
lazy/10
>>54004265
that shit looks like it would take a ton of data and processing, just figure out how long it would take to render/generate and be patient or make your own optimized version
>>54004810
But you can get better ones from aliexpress.
Hey, everyone, look at that cute cat. He is comfy and very happy but if you don't start programming in Racket from now, he will die at sunset. Do you really want to become a cat murderer for the rest of you life, anon ?
