Old thread: What are you working on, Jow Forums?
/dpt/ - Daily Programming Thread
roll
OP featuring young children has reached a new level
this is a CHINESE cartoons website
can I get uhhh
side projects
updating this porn highlight creator to include personal tastes
can i patent things pseudonymously, or patent via an LLC or something i wonder
dumb frogposter
dumb frogposter
>he's triggered by frogs, the face of the four chins
...
stupid frogposter go back to reddi
t
be comfy pls reee
I made an OCR dictionary. Could someone build some binaries for Windows?
shilling my free program as well
>Literature
>Carmen
Great resource, learned that algorithm design is a wild bird no one can tame from it
STOP WIGGLING THE FUCKING CURSOR YOU STUPID AUTIST
Studying for an logic exam in 3 days.
Fml how can anybody enjoy this slog?
Natural deduction and hoare logic are okay but fuck induction and by extension DFA proofs.
WRIGGLE LIKE A FUCKING EEL
WRIGGLE LIKE A FUCKING EEL
WRIGGLE LIKE A FUCKING EEL
Do go to school tomorrow
at least you can use this to flex on post enlightenment liberals
You can't spell spell without spell.
You can't spell word without letters
I just remembered I didn't update the Debian release yet so if someone is trying that out I think you also need ImageMagick to make it work.
So funny. That's why nobody will ever notice your stupid half-assed "app"
I know I'm no comedian but the program is a little bit more than half-assed, although the description may make it look otherwise. Only major problems are the SymSpell implementation which should definitely be paged and word boundaries on Asian languages.
I should probably get a better video, I think this one doesn't properly show that it works on all text and not just monospaced latin text.
maybe this isn't the right hobby for you, sweatie :)
JUST STOP MOVING THE CURSOR LIKE A RETARDED KID ON SPEED
How does my phone's autocomplete work? it's better than most simple markov things I've seen
People wouldn't know where the action is at though. Have to point it out to them.
free software could have answered this
It's even harder when you move it like if Michael J Fox was playing fruit ninja
It spies on you and everyone else to build a profile of what you're likely to type.
I really like programming in non FPLs and all, but I damn miss currying.
well I want a software implementation, so I'm just wondering about options and what they might have done for prediction better than "probability next word given current word" though maybe that's all they did
Yeah, all right. If I do another video I'll use my mouse like a normal human being.
Rajesh, my son...
>free software
>work hard to make software for people
>give it for free
Knuth-Moriss-Pratt is the best, isn't it?
>give it for free
I don't give my free beer for free
depends on the text
Or
>Work together with a number of people
>Software keeps getting better thanks to large contributions
Elaborate
Rewl
For me, it's Chapel.
>Who are you quoting?
Whomst speech are you referring to?
post what you have made with it or shut the fuck up already.
>can't use libraries compiled with it
pretty sad desu
There's that one with a vector of the size of all different chars we fill during preprocessing, quicksearch it's called iirc. It can perform sub-linearly Omega(n_t / n_p) or something like that because during a mismatch and if following char isn't in pattern, it shifts n_p+1. But there are cases it performs O(n_t*n_p).
Damn I nearly forgot all my algs class.
chapel-lang.org
oh well it might be coming maybe next update but currently it's in sad state
Boyer-Moore
Still working on it. Here's my current progress. I have 4 threads on my machine.
task 0, 0
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, 15
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, 30
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, 45
task 2, fizzbuzz
task 3, fizzbuzz
task 0, fizzbuzz
task 1, fizzbuzz
rolling
Careful user. If you forget your algorithms you won't be able to adequately pass the algofag riddles that every interviewer will pose to you.
I still remember my secret O(1)
FizzBuzz
FizzBuzz doesn't have any parameters so all implementations are O(1)
that's the secret
I don't know about you but I didn't get induction, at all, until I learned that it's really the same thing as recursion.
>FizzBuzz doesn't have any parameters
But it does though, you have the max number untill your fizzbuzz and then stop
No, it's 100.
rohl
what've we got
trying to organise a javascript study group on discord
Im only intermediate level, so would appreciate someone better than me or other motivated people
Invite:
t8hyH2
also
>discord
>want to make a game
>tfw can't play games
>discord
just make a facebook messenger group
gaymers btfo
cringe
based
learning haskell (again)
made RPN calculator
how bad is my code?
>if length stack then if length stack
dude just use pattern matching
oh nice. lol im learning rust and decided to make a stack machine as a first exercise as well. pic related.
nice color theme
passing everything by pointer or just returning new struct?
Allocate a new struct and return a reference to that
112
everything by pointer
I would use Vec::new() instead of [].to_vec() since Vec::new() is a const fn
You can write new_store() without return, by omitting the semicolon:
fn new_store() -> STORE {
STORE {
Binary: Vec::new()
}
}
set_binary does not need self to be mut:
fn set_binary(self, bytecode: Vec) -> Self {
Self {
Binary: bytecode,
.. self
}
}
Use Self to make refactoring easier - same as writing "STORE" everywhere, but if you decide to rename STORE to something else, it's less work, since Self always refers to the class name, whatever that class name is.
>Self always refers to the class name, whatever that class name is.
ah i did not know this thanks.
its the standard dark theme on sublime iirc
Is there any graphical debugger for haskell?
tried that one in VS but it doesn't work if program contains IO
>haskell debugger
>it doesn't work if program contains IO
pass a pointer, let the user manage the memory himself. read uncle bob.
redpill me on java lambdas
They're just a shorter (and better) way of writing some anonymous classes
Are Javafags even capable of contemplating something that isn't a class?
>inb4 "an object"
any idea what technology/programming language should i learn? i already have some experience with c#, nodejs, html, css.
a-an interface
I don't mind the fact that Java lambdas are actually just classes that implement interfaces. It makes it possible to gracefully migrate to the new way of doing things.
Haskell ofc
Since I want to catch all the exceptions I just put nothrow in main. Trere are thousands of throwable functions hahaha. Does try/catching them everywhere come with a performance penalty?
That's pretty cool but please don't use subhuman language to program and don't place 9001 files in the root folder
No one is going to look at the source like this
>try to use python
>doesn't even have a fucking goto statement
>have to use flag variables, continue, and break
BRAVO
how do you get a bitmap font in vs?
>goto
use try catch
On any modern platform there is no performance penalty for entering a try block, but there is a huge cost on finding the correct handler.
Yeah, that, uh, works. Thanks.
based on your current stack (similar to mine) probably python (I use it a lot to data collect and to make APIs with node.js or .NET).
react.js if you want to learn more frontend.
it's screenshot from github.com
The crux of is that you only pay for exceptions if they're thrown. Otherwise, there's no overhead.
>tfw mmaping files
what's the threshold to decide whether to mmap or not?