/dpt/ - Daily Programming Thread

I'm working on porn

Seeing if I can do row polymorphism with Haskell

File: 1471105341622.gif (234KB, 379x440px) Image search: [Google]

234KB, 379x440px

Okay, so you guys think I should stick with C instead of learning Python?

What are some recommended books or videos I should be looking into for a beginner?

>>56129817
Mixing languages when you are a noob is bad stick to ONE

I read somewhere that you can determine the number of digits an integer takes up by taking log10(n) + 1

what's a log10?

>>56129817
Python is a better beginner language.

>>56129817
Yes, stick with C.
Switching languages left and right while you're still learning the basics of programming is going to fuck you over.

>>56129823
log10(n) = log(n)/log(10)

>>56129927
what's a log?

>>56129938
log a (a^b) = b
ln(e^x) = x
log (x) = log 10 (x)
ln(x) = log e (x)

log b (a) = log k (a) / log k (b)

>>56130002
what's a ln?

        QRect rect_ = event->rect();

        //calculate width/height and x/y for inner rect
        qint32 w;
        if(rect_.width() >= 100)
        {
            w = rect_.width() * .8;
        }
        else
        {
            w = rect_.width() * .7;
        }
        qint32 h = rect_.height() * .6;

        qint32 x = (rect_.width() - w)/2;
        qint32 y = (rect_.height() - h)/2;

        QRect inner_rect_(x,y,w,h);

is this too confusing to read/shitty?

>>56130007
i told you

>>56130013
w = rect_.width();
w *= (w >= 100) ? 0.8 : 0.7;

>>56129817
Fuckface

>>56130026
thanks so much friend, completely forgot about conditional operators

>>56130026
I'm not sure I quite understand your notation
ln is the exponent part?

>>56129816
cute :3

File: why.jpg (10KB, 304x228px) Image search: [Google]

10KB, 304x228px

>>56130078
>ln(e^x) = x
>ln is the exponent part?
Srsly? (not the other one btw)

>>56130185
cringe

>>56130078
>>56130190
Also literally “man log10” gives you a good enough lead

>>56130207
:3

>>56130223
d:^)

>>56129823
>>56130214
>>56130207
Also log10(n) + 1 isn't equal to the number of digits; floor(log10(n)) + 1 does.

Anyone know a neat java (android) crypto library that you can also use on python? encrypt in java and decrypt in python

I've found some but maybe some of you already tried and liked one of them any suggestions?

>>56130254
Cool trick

>>56129785
anime isnt tech, stop trying to make this a thing.

File: users.png (81KB, 1169x876px) Image search: [Google]

81KB, 1169x876px

>>56130260
>Anyone know a neat java (android) crypto library that you can also use on python? encrypt in java and decrypt in python
A crypto library doesn't entail a specific form of encryption; it is entirely possible for one library to decrypt messages from another perfectly, provided you used the same *algorithm* (alias cipher maybe) on both ends. You should stop thinking that everything that is produced by one piece of software is locked to it for eternity and FREE THE INFORMATION!

>>56130260
>java (android) crypto library that you can also use on python
Just pick any Java library that implements a particular algorithm.

Use any Python library that supports the same algorithm.

>>56129817
K&R

>>56130298
>stop trying to make this a thing
Anon, this has literally LITERALLY been a thing for over 4 years now.

How new are you?

How do you guys estimate how much time a project will take?

Is there some books or other resources that have accords or something I can use to estimate how much time I should set aside for certain projects/tasks?

>>56130342
gantt charts

>>56129785
https://github.com/golang/go/blob/659dd4f1d7ed4e040d32346fa18c4ae3311ed81a/src/cmd/compile/internal/gc/lex.go#L604
https://play.golang.org/p/sI0h9sUZw-

>>56130342
>How do you guys estimate how much time a project will take?
By basing it on how long it took you to do it last time, or to solve a similar problem.

>Is there some books or other resources that have accords or something I can use to estimate how much time I should set aside for certain projects/tasks?
Yes, there are books that will tailor a time estimate to all of the factors of what comprises your work ethic and the total knowledge that you and your team have gained over the course of your lives.

Even the most simple projects are going to vary wildly by deliverable constraints, and even the most innocuous constraints can cause an additional 50+ hours to an inexperienced team.

>>56130342
Estimate how long it took you before and then double it.

Things always go wrong.

>>56130342
Estimate how long it took you before and then double it.

Things always go wrong.

>>56129785
>What are you working on, /g/?
im a retarded newbie trying to figure out where to start

File: odoo.png (5KB, 800x600px) Image search: [Google]

5KB, 800x600px

>>56130384
Mandude what the FUCK? Seriously what the FUCK? Why would you do this kind of shit? They even added a POOP EMOJI to the stupidest easter egg that ever was! Go is definitely the worst language ever made.

>>56130492
learn Go.

>>56130503
>They even added a POOP EMOJI
It's a metaphor

File: goat.gif (8KB, 300x165px) Image search: [Google]

8KB, 300x165px

HEY GEE got a challenge for ya! Translate this into Snek so that a local variable shadows a global/nonlocal variable while depending on its value! (Lua btw)

local val
function lel(inc)
    local val = val + inc
    print(val)
end
val = 1
lel(3)
val = 2
lel(2)
val = 3
lel(1)

three hours ago I believed it had to be purely impossible, but on second thought it may be if you put your sanity aside. GLuck!

File: 1466577279086.png (71KB, 780x530px) Image search: [Google]

71KB, 780x530px

>>56130727
>snek
No

>>56130299
>>56130312
Thank you

>>56130811
What's the issue here?

ddaa's comments seem reasonable.

>>56130811
>Python function call UGLY
>Python code

File: bananaman.jpg (232KB, 1536x1158px) Image search: [Google]

232KB, 1536x1158px

>>56130860
It's an unironic and reasonable remark that highlights the insanity that is Snek. What the fuck is the language itself for if you have to migglemac around endlessly in order for your incantation to cause some C code to loop instead of just looping? That lang isn't even Turing complete at this point.

Is there anon here with whom i've talked yesterday about recursion and iteration using explicit stack?

>>56130811
>try/except
What a meme language. You shouldn't even do simple conditionals and loops within the language because it's too slow, just tack on exception handling everywhere because nonlocality is more pythonic™.

Can anyone who's been working on C++ projects post their code (if they want to) so I can learn how to read code, please?

>>56131677
what is github

>>56131677

#include <iostream>
using namespace std;

main()
{
    cout << "Hello, world" << endl;
    return 0;
}

 file.write(u'\U+2005' + text) 

gives me this error:
SyntaxError: (unicode error) 'unicodeescape' codec can't decode bytes in position 0-1: truncated \UXXXXXXXX escape

I have no fucking clue how Python works with unicode, what am I doing wrong?

Hey do you structure tests? In python I make a a different test module `test_module_name.py` for each module I want to test, but I don't know when should I make a class and what methods should go into a class? Should I make a class for each function tested and the methods of the class test various cases for the same function? Wouldn't that end up in having too many classes? Should I just ditch classes altogheter? Using py.test if it matters

Hello, does anyone have any good youtube tutorial recommendations to learn C++?

File: e4160f2028.png (5KB, 502x78px) Image search: [Google]

5KB, 502x78px

>>56129785
/g/entlemans, I'm using Visual Studio and whenever I'm editing a single .cs file (outside any project) the error messages doesn't appears for some reason. In the left side of my image you can see doing weird shit and not getting errors, and in the right it shows erros since the file it's in a project. How can I fix this?
>inb4 program inside a project
I need to work on several snippets and compile them from CLI.
>>56132444
Just use a book, tutorials are generally shit. (Except for some really specific concepts.)

>>56132444
>youtube tutorial
>to learn C++
why do people do this? can't you read a book or something?

>>56132744
i can't pay attention long enough when reading

>>56132185
if you're not using the class to represent something you could call an object, the class isn't doing much good. so probably best to ditch the per-function classes

Is there a PL out there that hasn't made disastrous feature choice? Is there one without fatal flaws?

>>56132763
Read 5 pages at a time. Then have a break or whatever. I do this and it helps. I always do this with 2-3 books, and eventually I'll have read the entire book without feeling like I was reading a lot.

>>56132787
>classes represent objects
Classes instantiate objects. Not sure why a dynamic language wants classes, though.

>>56131832
I thought it'd be neat to do it with /dpt/ projects.

If your mentality is "If it works then it doesn't need improvement." Then you're a very bad programmer.

>>56132981
Go is perfect

>>56133601
>interface{}
>a, err

>>56129817
They only believe that because they're retarded. Python is more flexible for general purpose. Unless you're writing drivers or a kernel or something then you really don't need C. Having fun wasting time.

>>56132744
>>56132743
The media hardly matters, but it's easier to learn from someone instructing you.

has anyone ever been self-employed?
who (and what keywords) did you target?
I know everything and was to make some money
C,C++,C#,SQL,Networking,OpenGL
Is it possible to do this without creating poorly thought out apps?

File: pls_respond_10.jpg (93KB, 500x500px) Image search: [Google]

93KB, 500x500px

>>56132743

>>56130330
>lying on the internet
Fuck off op

>>56134214

Visual Studio needs the project system to track which version of the .NET Framework you're targeting and which DLLs you have referenced. There is no way outside the project structure to do this and therefore, Intellsense and Code Suggestions will stop working or rather, not work for single files.

Yes, it sucks massive balls.

>>56134631
The weird thing is: Intellisense actually works for single files, I searched and found this: https://blogs.msdn.microsoft.com/vcblog/2015/04/29/single-file-intellisense-and-other-ide-improvements-in-vs2015/
The only thing that doesn't works are that, the IDE doesn't checks for syntax errors.
Can anyone who has installed VS Community 2015 check this?, Create a single .cs file and make a type to see if the editor can point it.

Damn, I'm going to install visual studio code.

>>56134722
and make a typo*

>>56134722

>C++

Yeah, good luck. It does seem that Visual Studio Code would probably be better.

>>56134747
Ooops, didn't knew that article was for Sepples. I'm using C#.

File: 1408830619293.jpg (3MB, 3504x2336px) Image search: [Google]

3MB, 3504x2336px

>>56134722
>visual studio code
Holy fucking shit, it's the worst editor I0ve tried in my entire life. It doesn't comes out with C# coloring by default, it's super slow, doesn't shows syntax errors, and that fucking thing spawns 6 threads.
I give up, whenever I'm going to edit a single file I'm going to do it inside a project.

>>56135102

It's based upon Electron so what are you exactly expecting from something that refuses to use a proper GUI?

>>56135123
I think discord uses that same thing and it doesn't sucks balls that badly.

>>56133399
we all know

I thought I'd just make a C++ project, it'd work out fine, it'd be great.
I didn't know it would segfault for literally no good reason.

>>56135144

It does and I refuse to use anything that isn't a web browser using Electron or Blink as a dependency of any kind.

Qt bloat is already ridiculous enough at 15 MBs of a basic GUI window if you statically link it. I don't need something that over 2x that size.

File: 13892242_1749468705293619_2362727606811283003_n.png (366KB, 720x720px) Image search: [Google]

366KB, 720x720px

Trying to get this bash shell parser working

IN="firsttext\nsecondtext"

IFS=$'\n' read -ra ARR <<< "$IN"
for i in "${ARR[@]}"; do
    echo $i
done

The output gives me:

firsttext

seco
dtext

I just want

firsttext
secondtext

What am I doing wrong? Very new to bash, sorry if stupid question

This thread is dead. I guess everyone gave up on programming.

File: output.webm (2MB, 260x462px) Image search: [Google]

2MB, 260x462px

>>56129785
Alright /g/, r8 my space shooter (it's still far from getting done)

It's completely written in Swift and Objective-C

> please don't hate on the ad, it will be a free app

>>56135606
Inconsistent art style/10

>>56135606
ehh... did you draw those icons by yourself

Is there a good reason to use Python3 over 2.7?

The latter is some 50% faster in my experience.

>>56135762
I want to know answer to this too, I'm not a python guy but I use it when necessary and it seems like there's just more external library support for 2 over 3 but I'm not sure which one I should really be using.

>>56135789
I just write python3-compatible python2 code to minimize the pain when python2 dies.

>>56135200
Well I see I lot of those people here.

>>56135718
yes, are you proud of me?

Can someone post one of those programming projects roll images? I need to get back into things, haven't written a line of code in months.

Why do some of you use C/C++ for general purpose? That's kind of ridiculous. Why wouldn't you use something like Python?

>>56135912
What do you consider general purpose? Give some examples.

>>56135324

bumping for this

>>56135911
http://better-dpt-roll.github.io/
For the image, search the archive.

>>56129816
whats row polymorphism?

File: v4.0.png (2MB, 3840x2160px) Image search: [Google]

2MB, 3840x2160px

>>56135911

>>56135967
polymorphism based on which fields ("rows") an object has, rather than an explicit type hierarchy

>>56135922
Almost anything other than drivers and kernel source code pretty much.

>>56135993
give me an example

is is just records, like

data Person = { name :: String
                       , dead :: Bool
                       }

because one can sepecify any order as long as all fields are given

>>56129785

Distributed systems with Elixir/Erlang.
The results? Here:

https://www.raskie.com#mememe

what to work on?

https://www.debian.org/devel/wnpp/orphaned

>>56136006
You couldn't even give 1 example, so your opinion is worthless.

>>56135987
roll

>>56136085
I could give hundreds retard. Unless you write drivers, kernel source code or other software that needs to physically interact with hardware then you don't need C. It's almost a complete waste of time.

>>56135897
yes big boy

>>56136147
>asked to give an example
>can't give an example
>asked again to give an example
>'I can give hundreds retard"

Once again, give 1 example.

>>56136196
Web browsers, back-ends for websites, image viewers, image manipulation programs, physics simulations
There, I picked a few random things from the category of "other than drivers and kernels"

>>56136074
explain

>>56135912
You know what Python libraries are written in right

>>56136196
You're the kind of people that ruined 4chins
>I'm right
>I'm right
>I'm right
>Oh... I guess I am wrong. I'm going to stop replying now because I got btfo

>>56136302
Yeah, They're written in C. Why do you think Python was created?

>>56136218
which package(s) to take over development for

>>56136313
>Yeah, They're written in C. Why do you think Python was created?
for retards like you

>>56136321
How long have you been programming in C? It is pretty much a waste of time because of the flexibility of Python. Just Cythonize everything. People who learn C as their first language and only use C are fucking retarded.

>>56136352
I mean, if they want to learn OS dev, that's not very retarded to do. But I learned C first, and have done most of my programming in C. I could definitely have done a lot more if I learned Python (or Java, eugh) earlier than I did. I do still think it's wise to learn C first though, just because it's simple, and not dynamically typed.

>>56136313
>Yeah, They're written in C. Why do you think Python was created?
To help people make prototypes? I don't see what you're arguing since you acknowledge that external libraries of python are written in C (or C++ like opencv), so at least something has to be written in C. Whoever writes it has to know C and that's what they program in.

>>56136352
but why cna't people just program in a language they want? And why are you implying that it's the only language they program in?

>>56136021
Something like

(psuedo haskell)
doThing :: {value::a} -> a
doThing obj = obj.value

data T1 = O1 { value :: Int }
data T2 = O2 { value :: Char }

doThing (O1 0) = 0
doThing (O2 'a') = 'a'

>>56136074

https://en.wikipedia.org/wiki/List_of_Unix_commands

Anything that brings debian a bit closer to being actually POSIX compliant, and can be improved on. Some of the stuff are deprecated so understandbly are dropped like the SCCS stuff which we have SVN, Git or Mercurial.

>>56136392
>>56136402
Don't get me wrong. C is a great language. I mostly stick to Python and C. With that being said though working with C can be pretty unnecessary. I'm not implying that everyone who uses C only uses C. I just feel like I've wasted a lot of time with C because there's Python. With how flexible it is it pretty much renders C almost useless to me. C is great but I would hate to use it all the time for general purpose.

>>56136454
For general scripting, don't use C.
For huge applications, don't use C.
For low level interfaces, use C.
Just weigh your options, really.

>>56136454
I only use C more than Python because I am more familiar with it. I'll honestly just not remember easily enough how to do stuff in Python, and using C (or C++) is just easier, only because I know it better.

I need to force myself to do some projects in only Python so that I become familiar with it better.

>>56136021
imagine you've got two records

data X = x { name :: [String], age :: Natural }
data Y = y { name :: [String], age :: Natural, address :: String }
(this wouldn't compile because there are two lots of name and age )

A row polymorphic function that operates on x could also operate on,
e.g.
print = show name ++ ":" ++ show age

similarly, you've got a variant
data N = A | B | C
data M = A | B | C | D | E
A row polymorphic function taking an M could also take an N

>>56136475
Yeah. I just don't understand why some people only use C for everything. (It's perfectly fine if they're still learning.) If my job required me to use C for everything I probably kill myself eventually.

>>56136509
That's understandable. I had some trouble coming back to Python after working with C for a while. I just feel like Python might suit some people more for the task. Python and C also work great together so you may want to consider that.

>>56136512
>(this wouldn't compile because there are two lots of name and age )
this was fixed in GHC 8 with duplicateRecordFields extension

>>56135993
>polymorphism based on which fields ("rows") an object has, rather than an explicit type hierarchy
why is it "row polymorphism" if the convention in every other theory (e.g. database theory) is that object records are rows and fields are columns (hence object-major storage is "row storage", and field-major storage is "column storage", i.e. column orientation databases).

please, please tell me that the PL guys have not inverted the most basic conventions of a decades-old field of theory

>>56130330
Stop posting homo

>>56136603
That's actually really funny. Yeah that would suck.

>>56136603
it operates on many records based on what fields they have

so it's polymorphic on records

>>56135912
modern C++ has enough functional iterator conventions to make many data processing applications easy to write, and will slaughter python on performance, and linking it with C libraries requires much less boilerplate

it can be hard to finish a serious project in python because of performance issues. one valid approach is to prototype and then rewrite parts in C, but another valid approach is to be familiar with modern C++ libraries. C++ has lambda, higher order functions, iterator sugar, etc.

>>56136512
>A row polymorphic function taking an M could also take an N
not necessarily...

data N' = A | B | C | D
f :: [> N < N'] -> a

in this case, f would accept N and N' but not M since the types of N and N' are "less than or equal to" N' but "greater than or equal to N". That's not true for M (since it's "greater than" N' because it has an extra value)

>>56136637
>it operates on many records based on what fields they have
>so it's polymorphic on records
that's acceptable, the person i quoted said "fields (rows)" and i almost fainted

>>56135967
Haskell people love to use big words to sound smart.

>>56136653
what?

N = A | B | C
M = A | B | C | D | E
this is isomorphic to

M = ABC N | D | E
so to call an M function with an N, you just compose the original function with ABC

>>56136653
>>56136675

Oh I think we misunderstood each other
I said a function taking an M could take an N, I didn't mean to imply the reverse

>>56136313
>can't give an example

How's it feel to know everybody else knows you're retarded?

>>56136637
You missed the point of his post, anon

tables are like this:

_| Object | field A | field B | field C
   Stan     12        106       99
   //.. etc

row polymorphism should be called column polymorphism because in the table, rows are for each object, and each COLUMN is for each field, i.e. if objects have the same COLUMNS then [...]

>>56135762
there is some neat shit in the pipeline like the types library that may save some pain

>>56130007
A log taken from free-range trees and not trees grown in a factory farm.

>>56136674
Haskell doesn't have row polymorphism, most languages do in some form

>>56136699
It's polymorphic on records, depending what fields they have
Polymorphism - many forms
many forms of record

>>56130342
Always estimate that things will take an infinitely long time. That way your boss can never get mad at you for being late and whenever you're done you'll get praise for finishing early.

>>56136403
ok oh, so make the record fields private?
or make them work based on contexts
kind of like read

>>56136719
We're talking about the term "row polymorphism", why did you feel the need to tell me what it is?

>>56136699
no, it's polymorphic over the rows (i.e. any row gets excepted regardless of whether it has all the columns)
for example

_| Type | field A | field B | field C | field D
   ABC  | 12      | 106     | 99      | XXXXXXX
   ABCD | 3       | 29      | 17      | 24

>>56136719
>most languages do in some form
I'd say very few do. Row polymorphism is technically distinct from subtyping (even structural subtyping). One of the requirements for row polymorphism is type inference (with some type system more expressive than HM). So even though you can have structural subtyping in Javascript or Python or something, you don't have row polymorphism technically

>>56136791
>>why is this thing called X?
>X is called X because it does X

>>56136803
C++:

template <typename T>
auto func(T& x) { x.field(); }

+ subtyping enables row polymorphism, even if row polymorphism doesn't have to be subtyping

>>56136830
i fucking wish C++ field names were open to template type inspection, oh my fucking god. i would write the sickest in-memory column DB

>>56136830
C++ having it does not make most languages have it. also, im not sure if that would technically qualify as row polymorphism there or as structural subtyping, since it still requires you to define the types being inferred ahead of time. row polymorphism still has to solve for types that have not been explicitly defined (i.e. anonymous types).

>>56136913
C++ types only exist at compile time, inheritance is the only runtime polymorphism it has

>>56136934
which would work well, since row polymorphism can emulate inheritance

>OOP sucks
>I'll use an FP language
>I'll implement OOP concepts in this language because I like them

>>56136974
>OOP concepts

This is bait, right?

There's no way this can't be bait, right?

>>56136982
Polymorphism is an OOP Concept. It's in Java.

I don't know If this is the correct thread or If I shoudl post this at all on /g/...
But I really don't know what to do with my (future)career.
I have learned a little bit of C, C++, java, asm, C#.
I learned some in HS other is self taught, but nothing really in-dept.
I don't Know what to canalize my learning on and what should I specialize in, please help!

>>56136934
>>56136964
a lot of C++ game engines essentially implement this idea at runtime in their entity systems. "entities" are rows in a column-stored table with sparse columns, and stream operators apply to all entities that match a type defined by the presence of a subset of columns, which may be assembled ad hoc

>>56135912
Python's syntax is just awful IMO, so I can't stand to use it for even the simplest of tasks.

>>56137008
Just pick one of the languages (C, C++, Java, C#) and really learn it. Like, really learn it well. I recommend C++ or Java because there are a *lot* of libraries available to do cool things with. Basically, read a book that teaches programming using the language, and then make projects. Console based or simple GUI projects are fine, then find cool libraries and make new projects using them, then find what you're really interested in

>>56137008
Why would you want some /g/erms decide your life? Just look into what you like, and specialize on it. On of my biggest mistakes was keep changing between languages and fields.

@56136990
(U)

>>56137026

def main():
    # ...

if __name__ == '__main__':
    main()

>>56137026
imo all python needs is brackets, types, and performance

we might eventually get the latter 2 but never the former, unfortunately

>>56137050
it's stupid but it's not substantially more stupid than the magic main entry point of other languages

>>56136852
It's a shame. Maybe they'll propose it in the C++ 2099 standard.

>>56137062
it actually is.

>>56137008
what kind of computer stuff interests you the most? if you like a topic or think it sounds cool/intriguing just start studying it as much as you can. you don't have to ever complete anything in it, but you'll learn along the way if you keep trying something or studying it.
>>56137023
sounds pretty cool. do you have any example code (that's actually in use) that shows this?

>>56137055
Really the main issue for me is the lack of brackets. There's no rhythm to the code without them.

>>56137062
if __name__ == '__main__' is pretty magic dude.

>>56137036
>>56137035
Thanks for the replies.
I really don't have Idea of what I want, I mean i like programming since you are basically making your thoughts reality inside a computer(sorry if this sounds autistic).
But a firend of mine says that the faster I choose an specialization the better, so I am kind of worried about that.
Also if i can fall back into childhood dreams i guess I could same game dev, but that is retarded as pay is bad and you can't really make a living out of it, I guess as a hobby it might not be too bad...
Anyways can some one point me or tell me in their experience what are the most welcoming fields in this industry or a place where I can learn more about the different ways I could go?

>>56137055
Why don't you just use a different language?
Why does everyone want so many features added onto their favorite language when there exists better languages that already have the features?

>>56137103
That's a retarded complaint. You shouldn't complain that there is no rhythm, you should complain that the rhythm is forced.

#include <stdio.h>
int
main(
int argc, char **argv)
{ puts("This is valid C code");
return 0;}

def main(args):
    print("This is Python code, and I couldn'tve have spaced it out any other way.")

>>56137129
Great response

>>56137128
You don't need to choose a specialization. That's like asking a middle schooler what they want to do for the rest of their life.
I've been programming maybe 3 years or so and I don't know what I want to do. If you just learn and build, you will get a feeling for what you like.

>>56136852
>>56137063
it's not impossible to do at the call site

you can make a template that performs differently if a certain parameter fails to typecheck

#define hasMember(m,t) decltype(std::declval<##t##>(). ## m ## )
ifCompiles<hasMember(mymember, T), ifTrue, ifFalse>

>>56137167
#defines are a horrible hack compared to how great actual field-templates would be

>>56137165
...If you took away dynamic typing, significant whitespace and bytecode interpretation, Python would cease to exist. Also, brackets are overrated.

And isn't pypy blazing fast?

Guys is Lua worth learning?

What "real world" applications does it have?

File: 1469976671467.png (654KB, 2640x1836px) Image search: [Google]

654KB, 2640x1836px

>>56137128
Program things that you have fun making and can do solo. If you've got a project to do every week you'll become better. Specialization is silly, but you should be able to know your shit about at least one thing.

>>56137222
vidya games
it's a simple language, learn it

>>56137081
>sounds pretty cool. do you have any example code (that's actually in use) that shows this?
Unity and Unreal both use an entity-component data ~model~ and I know Unity uses C# generics to access entity collections by their component types. for Unreal I think it is a mutant flipper baby because mature engines need to support designer-facing runtime scripting languages, and this leads to a weird fusion of compile time and runtime component types.

there are a shitload of GDC talks and articles on entity-component design, but i think most systems like that don't survive contact with the sloppy realities of game design.

the main driver, aside from developer convenience (compared to ridiculously deep game actor type hierarchies), is the data bandwidth gained by sparse component column storage, since components are usually accessed by a game system all at once (i.e. geometric data for collision testing)

>>56137128
game development is fun if you treat it as a hobby. there's a lot to learn out of it - optimizing for performance, networking, lots about graphics programming, sound, etc. as well as just generally having applications for lots of data structures and algorithms that programmers tend to use regularly.
and there's definitely nothing wrong with specialization because everything overlaps and you'll only work hard if you're enjoying what you're making. if you like making games but you're forcing yourself to write web backends you're gonna get burnt out a lot faster and you'll just give up. but if you can tie in the server stuff to games (i.e. a game's server) then it's a lot more fun to work on because you have a clear goal in sight and you can actually have fun seeing the end product.

>>56137129
>Why don't you just use a different language?
because a language with the positive features of python + types does not exist

they added type annotations in recent versions of python3 so they seem to agree

>>56137278
>the positive features of python
i still dont get what these are

>>56137167
you can do template tricks with explicitly named members but you can't e.g. enumerate the members in something "like" a typelist, which prevents you from decomposing a struct by its member types

this is the reason that tuples need to exist and is the primary way in which C++ tuples differ from structs -- you can access the typelist of a tuple, but not a struct

>>56137117
yes, which is why it's about as stupid and not more or less. all program entry is fairly arbitrary

>>56129823
> asks /g/ instead of googling it

Serious question /dpt/:
Is c# my best chance at getting a job with game dev + emergency plan b stuff like pajeet work?

>>56137354
yes

>>56137290
python has a lot of good libraries and a few good language constructs for coupling libraries together, e.g. generators, which in a language like C++ are much more verbose

if i want to like, connect to a web page, submit a bunch of GET queries to it, parse a bunch of data out of the response and then feed that data into structured storage, in python that could very well be like 100 LOC or something stupid, and very legible. whereas in C++ you might be writing class DocumentIterator

>>56135211
Your initiation is now complete.

>>56137380
>lot of good libraries
Haskell has a lot of good libraries
>e.g. generators
JavaScript has generators. Racket has generators. C# has generators.
>connect to a web page, submit a bunch of GET queries to it, parse a bunch of data out of the response
node.js was designed to make this task easy. JavaScript also has brackets, and typescript has types. Maybe you're looking for JavaScript?

>>56137442
Haskell is a meme

>>56137463
A fun meme, though.

Can you recommend any good guides and libraries for getting into signal and image processing?

>>56137442
>Haskell has a lot of good libraries
i can't name a single discipline that maintains its gold standard libraries in Haskell. can you? (it's also a garbage language and i know multiple people who've dashed a year against the rocks failing to write a project in Haskell, but let's focus on libraries).

>JavaScript
JS is a fine language but AFAIK has very few scientific computing libraries, which are commonplace in python to the extent that matmult gets an infix operator. if you don't need access to those then yes, I think JS is a good alternative to Python

I'm making an entity-component system in Clojure that uses a sqlite database for all the data.

I know an ECS exists for Clojure, but I thought it'd be fun to make one myself. I am still learning the language.

File: 1471321331228s.jpg (2KB, 125x74px) Image search: [Google]

2KB, 125x74px

>>56137554
>JS is a fine language

>>56137512
this is, ironically, a good use of Python because there's a bunch of basic libraries like this that are easy to use

python image library (PIL)
scikit-image and scikit-learn
pywt (pywavelet)
and of course, numpy/scipy/matplotlib

these things will fail you when you need high performance but make it VERY easy to get started

What's up with guile scheme?

>>56137580
understand that when i say that i am acting collegial and mature on the outside but i am screaming on the inside

please someone post the javascript how to add numbers on stack overflow

One question. How do you determine the "size" of each constant. What I mean is, when are you supposed to use "double", "long double", "long int", in a program(C). I am a C beginner, I like it a lot so far, but that is something that has always bugged me.

>>56137587
still the best Scheme implementation

>>56137622
0/10

File: JUST.png (141KB, 1366x768px) Image search: [Google]

141KB, 1366x768px

JUST

>>56131832
kek

>>56137625
It's pretty fucked up and depends on the computer you're running the code on. And there's so many weird combinations.
I stick to just using "int" for most integers, "size_t" for array length, array index etc., "long long int" for extra wide integers, "double" for floating point numbers.
If I specifically need to know the # of bits, then do #include <stdint.h> and use [u]intXX_t
e.g.
int32_t for guarunteed 32 bit sign ints
uint8_t for bytes
uint64_t for hash codes

A lot of graphical libraries force you to use "float"s but if you aren't force to use float, always use double. You probably won't need to use "long double" ever.

What's pull request etiquette on Github? I forked and wrote an improvement to a small library. Should I just submit a pull request out of the blue or contact them first?

>>56137511
F# if you want a functional Functional language

>>56129785
Just wrote a generic binary tree in CuteSharp. Bruv I miss messing with C pointers.

>>56137625

General rule of thumb is to not deviate from the general rule of "int for integer numbers, double for floating point numbers" unless you can't fit a number, calculated or not, with your chosen type.

You have stdint.h if you need something with a defined bit width.

>>56137663
Thanks familia, I will take your advice into consideration. I can't wait to git gud with this language so that I can start doing cool stuff with it.

>>56137674
>Should I just submit a pull request
Yeah, do that.

>>56137628
What's so great about it? Honestly asking out of curiosity, I'd never heard of it until recently

>>56137677
>windows

>>56137554
Maybe ask why alnost no other language has libraries like semigroupoids, free, lens, mtl, etc.

They can't even express the concepts in most cases, let alone implement them well.

>>56137683
Aight, thanks man.

>>56137677
F# stands for first-order, its type system is crippled

>>56137686

Well when I first learned Github I submitted pull requests to add licenses to projects wihtout them and got a lot of hate so I thought maybe to contact first

Really stupid question, but can someone explain to me the concept of "kernels". New to linux, have a rough idea of what they are, but I don't truly understand it.

>>56137622
my favorite stack overflow of all time was a little more advanced
>Q: how do i list all the cycles in a graph?
>A: use tarjan's algorithm (250+ upvotes)

>>56137731
>I submitted pull requests to add licenses to projects
Well, desu, that doesn't helps that much to the project itself, some programmers doesn't care about licensing their software. (I don't care about licensing mine, prolly because it's shitty, but yeah.)

>>56137714
>Maybe ask why alnost no other language has libraries like semigroupoids, free, lens, mtl, etc.

No one cares about this shit.

>>56137714
>Maybe ask why alnost no other language has libraries like semigroupoids, free, lens, mtl, etc.
nobody except PL weenies care about language features in and of themselves. they want them only insofar as they (drastically) simplify problems in their application domain. if semigroupoids are so useful then why has no applied field switched to haskell for them?

so many haskell concepts are solutions in search of a problem. right now all the best problems in CS are algebraic and python's infix matmult is more useful in practice than any haskell abstraction (and it's not like matmult is a trivial problem, there's plenty of hard CS research behind that matmult)

>>56137731
Bit autistic lad

>>56137731
That is really fucking autistic dude. This has to be bait.

How do you debug/edit a program you already compiled? My system is currently gcc compiler.

>>56137677
>F#
>Functional

>>56137641
>windows 10

>>56137731
copyright adheres without a license. there's an implicit nonfree license on any code without an explicit one

>>56137822
>debug
use gdb
>edit
modify the source code and compile again

>>56137811
I can't speak for other players, but the company I used to work for used equivalent in Scala regularly. It was fascinating to see us gradually move from using the language as a better Java to using it as a worse Haskell. Now we're using generic transformer stacks like they're completely prosaic.

>>56137822
debug is gdb but a lot of people feel that testing ndn printf debugging are more efficient. i only use gdb for segfaults

>>56137822
You can edit it with a hex editor; hexedit is good enough

>>56137822
>>56137841
>>56137859
You can also use valgrind for debugging memory problems

>>56137811
Blub paradox

>>56137857
scala is a gateway drug and insofar as haskell is a research language its proponents should be happy that its having that kind of impact on other languages. but not every research idea is good

the value add of scala over haskell in terms of compatibility is fairly obvious and we both know that was the reason people were willing to take the plunge with scala

>>56137752
That's incorrect, though.

>>56137722
less crippled than ocaml

>>56137897
paul graham is a "tech entrepreneur" and tech entrepreneurs are a curse

>>56137900
Scala has interoperability with Java, true, but the language itself is a frustrating inconsistent mess, and the compiler is sooooo sloooooooooooow. Unlike F# though it at least allows you to think higher-order thoughts, even if you have to do a lot of manual plumbing that wouldn't be necessary in Haskell.

>>56137923
Not an argument

>>56137918
Both are pretty bad desu

At least F# can do multithreading, I guess

Anyone know about 3DBuzz? I have C++ and OpenGL video tutorials from 2004/2005. Would they still be useful?

>>56137937
invoking a meme from someone else's blog post, without commentary, is also not an argument

>>56137945
F# has computation expressions as well as active patterns

>>56137945
according to /dpt/, ocaml is better than haskell
so if ocaml is bad, then...

>>56137973
and actual type inference over scala

File: 1470955843833.png (389KB, 756x678px) Image search: [Google]

389KB, 756x678px

>>56137977
>according to /dpt/, ocaml is better than haskell

>"hey what are you doing?"
>"i'm posting in the daily programming general"
>"are you helping people learn how to program?"
>"no i'm arguing about why haskell is shit"

>>56137977
/dpt/ thinks Python is good right now, that should tell you everything you need to know about the merits of /dpt/'s opinions.

>>56137973
Cute gimmick, shame it's so limited

>>56138007
Who's this qt, I wanna hit the panda.

>>56138007
>>56138018
their arguments are that ocaml has some sort of module system and that it has no lazyness
they say that lazyness killed haskell, and because ocaml has no lazyness and a great module system, its better

>let's design a new language and not give it HKTs!
What the fuck are you doing?

>>56138018
less limited than Scala for-comprehensions

>>56138063
>HKTs
???

>>56138061
Oh ffs, laziness is one of Haskell's strengths

OCaml's module system definitely isn't perfect, check out Ed Kmett's talk on typeclasses

>>56138063
not immensely useful for my problem domain

>>56138080
Higher-kinded types

File: happy wolfram.jpg (81KB, 501x504px) Image search: [Google]

81KB, 501x504px

>>56137641
>mathematica
Dump that bloated trash in the trash and get Maxima.

>>56138087
Disagree, they make my life a lot easier at work.

>>56138091
I notice this is a thing in the haskell world
GADTs,
HKTs

>>56138061
{-# LANGUAGE Strict #-}

Modules can be emulated in Haskell

>>56138099
>>56138087
Sorry. Misread your post as "not useful for any problem domain"

>>56138100
I'm on my phone, typing out in full is tedious

>>56138101
Hiss

>>56138101
i know nothing more about ocaml
how does its module system work that make it so good

what are some commercial or scientific projects using haskell?

>>56138117
there's a page documenting such

>>56135606
Make the ship animate during the last hit. The ship should kinda freeze and have a couple frames of it blowing up. Also make the hearts more responsive to damage. Overall just add more animations I guess.

>>56138128
I don't know why people ask this, it must have been asked tens of thousands of times now.

>>56138134
it just adds to the notion that haskell is a meme and that there's not real world applications written with it

>>56138128
i wasn't really motivated by said page last time i checked it, it seemed like a melange of startups and one-ofs

i'm trying to find someone successful doing heavy lifting

>>56138148
GHCI is nice for command line stuff

>>56138148
whether its adopted for use doesn't matter
what matters is that there are libraries made for such
I belive hackage has a great deal of packages which include many data science libraries

>>56138155
>Are there any good projects using language X?
>it's REPL is a good project
Shit dude.

>>56138116
>modules are typed
You can use 2 modules of the same type interchangeably, and use them against in interface (like OOP interfaces)
>modules can be parameterized over modules
you can define a module that you construct based on another module, for example a hash table that works with a module containing a hash function + a type:

module type Hashable = sig
  type t (* The type of the hashable value *)
  val hash : t -> byte array (* Hash the value into a byte array *)
end

then you can define a module with that as:

module HashTable (H : Hashable) = struct
  ... (* Define the hash table and functions on the hash table based on the module H *)
end

this is really nice because you can use it to reduce boilerplate like that because if you already have some functions defined you can derive an entire module based on that (Cohttp derives the networking boilerplate to implement the HTTP protocol based on the networking module you give it. so you can give it a module with like 5 functions and it's able to construct a whole HTTP library with that). it's also nice for configuration (my telegram lib uses a module to configure your bot, MirageOS uses module parameters to decide which library to use in different roles (networking, file systems, etc.) based on the compiler target and the requirements of your project).
>modules are first-class
you can pass modules as normal function arguments, which is effectively the same thing as explicitly choosing a typeclass to use in Haskell (like you could pick out one of the hash functions for your hashtable by passing a module directly to that function)
plus, you can just dynamically pick out what module to use because modules are typed

>>56138170
>I never use the command line

>>56138167
i'm looking for projects like Microsoft Bond, i.e. modest-scale production systems in which Haskell seems to be more than glue, in order to get an idea of what a Haskell system looks like

>>56138203
https://wiki.haskell.org/Haskell_in_industry

Guys, I want to practice my math to be able to be a better programmer(I know it's not required, but I think learning mathematics concepts will help my coding skills greatly). Do you guys happen to have some links with some useful math problems with explanation and answer? Thanks in advance.

>>56138192
Do you not fucking get it? It's like saying:
>>What are some good projects done in C?
>The C compiler is a good project.

>>56138228
>I never use the command line

>>56138212
i've read through that page twice. most of the entries are not what i'm looking for. a lot of them feel like advertising, where they're taking advantage of an internal one-of Haskell project as an excuse to put their company on a high-traffic page visited by advanced programmers. i used the specific example of Microsoft Bond because it links directly to a github full of production system code

>>56138155
Lisp REPLs are far better, with interactive debugging and access to documentation.

* (describe #'mapcar)

#<FUNCTION MAPCAR>
  [compiled function]

Lambda-list: (FUNCTION LIST &REST MORE-LISTS)
Declared type: (FUNCTION ((OR FUNCTION SYMBOL) LIST &REST LIST)
                (VALUES LIST &OPTIONAL))
Derived type: (FUNCTION ((OR FUNCTION SYMBOL) LIST &REST T)
               (VALUES LIST &OPTIONAL))
Documentation:
  Apply FUNCTION to successive elements of LIST. Return list of FUNCTION
     return values.
Known attributes: call
Source file: SYS:SRC;CODE;LIST.LISP

vs.

Prelude> :info map
map :: (a -> b) -> [a] -> [b]     -- Defined in ‘GHC.Base’

>>56138223
the most useful math for a programmer, by a lot, is linear algebra. MIT opencourseware, khan academy, and probably any other online course platform will have a course

>>56138184
I see.
I'm not very versed in OOP or in using typeclasses.
But what I get from your post is that this module stuff is very flexible and useful.
I'd have to have more experience to fully understand everything you tried to explane.

Unrelated, but how is the gui and networking ecosystem in ocaml?

>>56138255
this is unfair, since one could use hoogle and hackage to get more info from libraries and functions

File: golden lel 3d.jpg (128KB, 776x678px) Image search: [Google]

128KB, 776x678px

>>56138255
>two useless languages competing over REPLs because they have nothing else

>>56138274
Thanks familia, gonna use khan academy, I have heard is quite good.

>>56138250
http://cufp.galois.com/2007/abstracts.html#CyrilSchmidt

>>56138255
>Lisp REPLs are far better
No they aren't

>access to documentation
That's a hindsight of GHCI, but you should really know what map and a lot of prelude functions do. Often you can tell from the type.

Obviously in some cases like libraries that's not as clear, but you could easily have a browser in the background

You could even just write a function that takes a string and gives back the hackage documentation, and then just call it in GHCI

>>56138293
REPLs are incredible

>>56138303
why did you reply to him, don't you know he's a troll?
they thrive on (you)s

>>56138287
You can do the same with CLHS. But accessing documentation with a single command is far more useful than digging through the Internet.

>>56138303
Just used map as an illustrative example, it could be any other function.

why is floor() and ceil() not called rounddn and roundup

>>56138317
Because that's not what they do.

>>56138317
because those names are more autistic

also, every length function should be called len

>>56138293
>I never use the command line

>>56136646
>C++ compiling is shit
>C++ comes with even more unnecessary crap than C

Then you don't know much about Python because it has great performance if used properly. Never heard of Cython?

>>56138303
>CUFP
cool conference that i'd never heard of, thanks

>>56138303
>>56138255
>>56138315

With this package
https://hackage.haskell.org/package/hoogle-5.0.2/docs/Hoogle.html

you can do hoogle searches (a search on hackage)


>>56138317
Floor = floor
Ceil = ceiling

Convention

>>56138313
>complaining about trolls
Trolling is a art

>>56138345
https://wiki.haskell.org/GHC/GHCi#Hoogle

>>56138287
>Unrelated, but how is the gui and networking ecosystem in ocaml?
GUI is, i've gotta admit, pretty shit. at this point the best GUI library for OCaml is probably using Electron through one of the ocaml->js compilers. there's ports of TK, GTK, and Qt but they all seem pretty bad from the 5 minutes ive taken to look at them
networking is a lot better. Lwt and Async both make it really nice and there's already some pretty good libraries out there (although keep in mind that multicore still isn't in the mainline compiler, so no big networking libraries take advantage of it). there's also some great networking shit for MirageOS apparently but i haven't really messed with any of it yet
https://github.com/rizo/awesome-ocaml#networking and https://github.com/rizo/awesome-ocaml#graphics are worth looking at

>>56138345
Still inferior to offline, fast access to self-documenting code.

>>56138372
It can access offline documentaiton

>>56138317
Ask Kenneth E. Iverson

>>56136697
One anon beat me to it. Still can't accept you're incompetent? Games, science, mathematics, web frameworks etc. Jesus Christ, I thought you would've wrapped your fucking head around it by now.

>>56138372
And lisp code is FAR from self documenting

>>56138352
>an art
ftfy

>>56137026
Now is it aweful? Dot notation is flexible, comfy and pretty amazing actually.

>>56138352
saying "language X is shit" isn't good trolling
>>56138367
danm it, both ocaml and haskell don't have good gui libs
is there any fp language that has a good gui lib

>>56138342
ok look, here's my gold test: when i call sort from python with an arbitrary comparator or key lambda function, does cython inline the comparator/key into the sort code, or does it pass a function pointer?

if the answer is the former then i'm very impressed, but if the answer is the latter then that's why i use C++

>>56138352
>>56138403
samefig

NEW THREAD!!

>>56138433
>>56138433

>>56138411
>danm it, both ocaml and haskell don't have good gui libs
*neither ocaml nor haskell have good gui libs

can someone tell me what this lisp code is doing

    (cond 
        ((typep is_atom 'atom) (push is_atom list_of_atoms)) 
        (t (loop for x in is_atom appending 
            (some_func x)) 
            ) 
        )

>>56138437
i hate you so much

NEW THREAD: >>56138414

>>56138446
No, what does it do?

>>56138411
F# i guess?

>>56138475
I don't like f# just because its a microsoft product

>>56138389
ML like types are so self documenting, in fact, that often the strict implementation of a function has only a few implementations

e.g.

id :: a -> a
-- has only 1 strict implementation
id x = x

bool :: Bool -> a -> a -> a
-- has only 4 possible strict implementations
bool x a b = if x then a else b
bool x a b = if x then b else a
...

instance Applicative ((->) r) where
  pure :: r -> a -> a
  -- one implementation
  pure x y = y

  (<*>) :: (r -> a -> b) -> (r -> a) -> r -> b
  -- suprisingly complex type signature, only one implementation
  (<*>) f g x = f x (g x)

>>56137345
>google sent me to g

>>56138415
I couldn't manage to make it pass a function pointer but there's probably a way.

>>56129816
You can do with a) a class that resolves TypeLits.Symbols to lenses b) heterogeneous lists labelled by Symbols as in Vinyl

>>56137962
lazyfoo

File: magic.png (36KB, 707x394px) Image search: [Google]

36KB, 707x394px

>>56139186
Yeah, I wanted to do it but I gave because coproducts are awkward (I was having troubles with type equality, I think it was a GHC problem)

I wanted to see if you could do it a non-Vinyl way using the new extensions, which allow you to do

x :: Product '[Int, Bool, Double]

>>56137859
>i only use gdb for segfaults
valgrind or fsanitize=address is better for that.
gdb is completely useless trash.

What are the most important things to take note of when trying to write a web-crawler?