« first day (3029 days earlier)      last day (1814 days later) » 

2:06 AM
@Downgoat How difficult do you think it would be to compile JavaScript to VSL? I have a brownfield JavaScript application that uses JS almost like it's C. I'm thinking of rewriting the actual hot computations in highly optimized VSL
Because if the blurb on the README is true, VSL sounds very very cool
 
@noɥʇʎԀʎzɐɹƆ hmm biggest issue would be typing, but there is already in fact a bindgen that (if oyu specify types of methods with a WebIDL document) you can use JS with VSL
 
I actually will contribute to VSL and use it if what it says is true is true
@Downgoat Hmm, explain? What would the steps be to compile JS to VSL?
I can definitely specify most types for it, but I want to retain the flexibility (other developers on the project who don't know javascript may not be able to type their code right)
What if I plugged the VSL backend into GraalVM I can run VSL with a JIT that interoperates with JS?? (That's a crazy idea, I would prefer to just get a reliable ECMAScript3->VSL compiler)
(Also where are the download binaries? The DNS doesn't seem to work for 'em)
 
@noɥʇʎԀʎzɐɹƆ As in the VSL WASM target will be able to interop with JavaScript (experimental DOM apis for vsl)
 
Hmmm how fast is the interop?
(Can I use pass-by-value or something as fast as it in VSL?)
 
interop is very low-overhead since VSL runs in WASM so almost bare metal
 
2:13 AM
does the LLVM target exist?
 
VSL doesn't actually pass objects it passes integers to and from JS that refer to objects which bolsters speed
 
~~create a wasm->llvm retargeting compiler and run vsl on the graalvm jit~~
 
@noɥʇʎԀʎzɐɹƆ as in a VSL->LLVM? That's what it does right now
 
cool
Btw, my git tip for you is: Use GitUp and commit every time you have a meaningful change, branch locally to experiment, cherry pick to get commits from one branch onto another, rebase to make git think the commit was always part of the history, and rewrite and split to clean your commit history
But very, very cool, sir
I'm part of a community that builds the game Arras.io. I can... look into mobilizing some developer resources to help with VSL stuff if you gave a bucket list of things we could do to help.
 
@noɥʇʎԀʎzɐɹƆ thats a good idea :P VSL commit history is a mess since I get so many features stuck in one branch and don't want to smush them all together
 
2:16 AM
@Downgoat yessssss Git is amazing when you use it the right way
I hated git until I read this ONE blog post and started deliberately practicing its method
In the next iteration of Arras' infrastructure/deploment system, Git will become my CI
 
@noɥʇʎԀʎzɐɹƆ For VSL? Current things on the to do list are dynamic dispatch for interfaces, optional types, and lambdas
@noɥʇʎԀʎzɐɹƆ git hooks ftw
 
@Downgoat yeah
push to release a random playtesting update (ssh pubkeys for access control); restarting the game server every time, pull from central master to update, updates on automatic arena close
@Downgoat Hmm, what's on the todo list for web idl?
I wonder what WASM's performance is on the serverside
I assume VSL isn't memory heavy
 
@noɥʇʎԀʎzɐɹƆ For WebIDL the main difficulty is actually finding WebIDLs that are standards compliant
you can find type definitions for chrome, firefox, etc but you want one thats cross platform
 
Hmmm link to example please?
 
as I work on the typescript interface for VSL hopefully that will make this easier
 
2:21 AM
In GraalVM, the C<-->JS interop is fast enough that if I reimplement a JS class in C, I increase the application's peak throughput by 50%. (eq. to a 33% cpu usage reduction)
 
(However, GraalVM 3.6xes memory usage compared to V8. And setting the heapmax does not help)
GraalVM does have overhead, but I enabled an experimental "inlining across language boundaries" and that helped
Can you please upload me a macos binary?
I'm interested in seeing where we can take this, I'm sure using VSL in production will only make it better :)
@Downgoat The docs link is broken help
How fast is compilation speed? Is there a "shortcut" compilation speed we can use to compile a debug build faster?
 
@noɥʇʎԀʎzɐɹƆ VSL has a binary format (.vslc files) which stores build information which means if you make a change pretty much only the file you change will be recompiled we can load everything else from the .vslc files
but as always -O0 results in faster comp speeds
@noɥʇʎԀʎzɐɹƆ hmm cant do binary because bundling LLVM would be complex, maybe somewhere down the line
 
2:46 AM
@Quintec can I have access to the fake cicada wiki?
 
 
2 hours later…
4:37 AM
got the new vsl docgen working :D
(well for basic things like classees)
 
4:59 AM
@Downgoat Ooh, packed pit Booleans, cool.
 
5:17 AM
@Downgoat I thought that said "Boo!"
 
6:16 AM
I only just now discovered f'{expr}' format strings in python <_<
 
6:43 AM
@betseg you should. doing some sport is healthy for people who sit in their office chairs all day long
 
7:14 AM
@flawr lol
 
 
4 hours later…
11:37 AM
ugh, Windows, can you at least make a Task Manager that doesn't crash randomly?
 
@Neil How often do you use Task Manager‽ I'd much prefer the resources being spent on making a File Explorer that doesn't crash randomly.
 
I have to leave it open because the graphs only record history while it's open, so there's no point opening it after the event.
and I don't know when the event I want to observe is going to happen of course
 
@Neil And it just falls over without you touching it?
 
yes
you can tell a) because the system tray sprouts an extra activity icon b) the window size changes to as small as possible
 
12:36 PM
Perfmon?
 
0
A: Sandbox for Proposed Challenges

Kevin CruijssenGroup Integer by Originality code-golfnumberintegerarray-manipulation Introduction: I collect twisty puzzles. Most twisty puzzles are produced and sold by Chinese companies. Most well-known companies ask permission from puzzle designers to produce their designs and work together towards a prod...

 
1:11 PM
morning chaps
 
yo
 
0
Q: Parse a C++14 integer literal

someoneAccording to http://en.cppreference.com/w/cpp/language/integer_literal, integer literals consist of a decimal/hex/octal/binary literal and a optional integer suffix, that is obviously completely unnecessary, wastes precious bytes and is not used in this challenge. A decimal literal is a non-zero...

 
2:16 PM
Well that was annoying. Finally figured out why my laptop's fans kept spinning up despite doing nothing. Audio driver stayed at like 20% CPU usage when having a USB headset and docked
Grabbed new drivers and Bob's yer uncle
 
@Downgoat How does one run VSL? Does the current master build?
Universal kolmogorov complexity: The first two bytes indicate which of the 65536 most popular languages on PPCG. The rest are the source for that language.
 
2:36 PM
@noɥʇʎԀʎzɐɹƆ see the wiki and yesh
 
2:57 PM
Ha ha, forgot Npcap adds a loopback adapter.
 
3:20 PM
We should rename ourselves to be the "programming problems and code genies" stack exchange. People can come to us with their problems, and we work to fix their problems for them.
 
0
Q: Convert a number into "bracket" numerals

Ethan SlotaLike most challenges, score is in bytes, least wins. "Bracket" numerals are a type of numerals I made up (not for this challenge), they aren't really numerals. Basically you convert the number into base 6 then replace a 0 with (, a 1 with ), 2 for [, 3 for ], 4 for {, and 5 for }. Your score ge...

 
@PhiNotPi sounds like an odd way to spell "Homework Stack Exchange"
 
Website of the day: applesnail.net A 21-year-old website all about a particular type of aquatic snail. I miss the good ol' days when the internet was filled with honest sites that were just trying to spread knowledge.
6
 
hi all
I am stuck on some homework about minimum vertex cover
argh :)
 
3:39 PM
How to sort an array of doubles?
 
@PhiNotPi that's adorable
 
you mean without called a sorting function?
@PhiNotPi if you provide one of those a day/week that would be awesome!
 
@Anush it was a reference to the famous code trolling challenge
 
@PhiNotPi Oh sorry.. but the website looks real?
I still don't think I get the reference... :(
 
"array of doubles" was the reference
 
3:42 PM
could you spell it out in terms a dim 10 year old would understand?
 
The website is just an interesting website I found.
 
ok.. so what is the reference to the famous code trolling challenge?
it is indeed interesting!
 
oh I see!
not snail related at all :)
sorry for the confusion
I really like that we have this category in our awards! "Best mathematical insight: Triangular Lattice Points close to the Origin by xnor"
 
@PhiNotPi Ha ha
 
3:56 PM
Bogo Sort is the only true sort
 
Bogo sort is bogus sort
 
4:12 PM
@Mego I am trying to use this Fibonacci heap to beat the standard queue implementation but I can't, can you provide any insights. I'm using it in Dijkstra on a matrix godbolt.org/z/U1Q3ec
 
@Rick when you say beat, do you mean you mean you want a speed up?
 
correct
 
@Rick that's a little tricky. you are trying to distinguish O( e + v log v) from O(e log v+ v log v)
where the latter is what you get with a standard priority queue
have you implemented the priority queue based method and got some graphs to test it on?
 
@Anush Fibonacci heaps is supposed to be find-min Θ(1) delete O(log n)[a] insertΘ(1) decrease-keyΘ(1)[a] mergeΘ(1) amortized time
that should beat the queue implementation
 
@Rick well it is exactly what I wrote
O(e + vlog v) instead of O(elog v + vlog v)
 
4:20 PM
correct
 
ok so.. have you implemented the priority queue based method and got some graphs to test it on?
you will need graphs with enough edges, to start with
and large enough graphs as the constants are higher for fib heaps
 
@Anush I just used the standard queue lib and I have tested it. on the average queue is faster about 2x faster
 
what graphs are you testing it on?
I mean how many edges and nodes?
 
@Anush i am using it on Dijkstra to search a 2d matrix in the example I provide. that's a 33X48 matrix I have a starting point and an ending point
Dijkstra is greedy so I search all adjacent nodes to find the shortest path
 
how many edges and nodes?
 
4:30 PM
the number of edges is one less than the number of nodes
with exception of the 0 those are wall
 
ok so you are not really answering my question
can you give the actual numbers?
 
yea one sec
41X33 = 1353 nodes
every node will have an edge to every other node where it's
neighbor is a 1
 
how many edges exactly?
 
connected edges or just edges
it's a complete graph 1352
/2
914628 edges
 
4:50 PM
what is the /2 part?
 
@Anush screen switch meant to type that into the calculator
 
0
A: Sandbox for Proposed Challenges

MilkyWay90code-golfnumber To exponential digit growth and beyond! Challenge Given a base (which is less than 10) and a term, your mission (should you choose to accept it) is to: Make a number in base-10 which will be the base + 1. We will call this "the current number" Do the following the term number...

New challenge
Can you guys suggest improvements?
 
@Anush that should be enough edges for fib
 
1
Q: Gambler's Fallacy Dice

BeefsterThe gambler's fallacy is a cognitive bias where we mistakenly expect things that have occurred often to be less likely to occur in the future and things that have not occurred in a while to be more likely to happen soon. Your task is to implement a specific version of this. Challenge Explanation...

 
unless I said something wrong
 
5:01 PM
0
A: Sandbox for Proposed Challenges

MilkyWay90code-golfnumber To exponential digit growth and beyond! Challenge Given a base (which is less than 10) and a term, your mission (should you choose to accept it) is to: Make a number in base-10 which will be the base + 1. We will call this "the current number" Do the following the term number...

 
@dzaima Have you gotten around to neighborhoods > radius 2 for your cellular automation yet? Sorry, I can't help with this because I don't really understand advanced Processing.
@PhiNotPi Presenting Stack Overflow for Homework!
 
@MilkyWay90 I have played around with it a bit. Just pushed the modifications for extracting a radius variable R you can change
sad that nothing can happen in radius 1, as it's impossible to grow at all
 
5:32 PM
@dzaima Thank you! Excuse me while I go play around with radius 2147483647
But in all seriousness, this is really useful
@dzaima I like the error message (NONSENSE!) when the depth is 0
I can just imagine a programming language where a parenthesis mismatch would return an error File main.xx, line 3: Syntax Error: NoNsEnSe
 
6:33 PM
Man I'm so close to getting NuStack to a TIO-able point
 
the new answers to codegolf.stackexchange.com/questions/184743/… have restored my faith in PPCG :)
now I just need 4 upvotes :)
there really is a great community here
 
that question has -4 votes
 
@Rick it does! It was even worse before
 
I think Anush's question is the most non-meta controversial question on PPCG
 
@Rick back to your question. A well known issue with fib heaps is the constant factors
 
6:46 PM
+10 -13
@Anush When did you lose it?
 
@MilkyWay90 as in when did I lose all the points?
 
@Anush I mean your faith in PPCG
 
@MilkyWay90 oh I can tell you
@MilkyWay90 when lots of people got upset by my using a restricted time tag.
 
@Anush can you elaborate on the constant factor
 
Ah, I see
 
6:49 PM
@MilkyWay90 and when lots of people downvoted and even closed (for a bit) my question for no good reason, as far as I could see
@Rick big Oh complexity hides constant factors
 
@Anush It was similar to your previous question, but not a duplicate
 
@MilkyWay90 right but really not similar for reasons I explained in the question and is covered here codegolf.meta.stackexchange.com/questions/14819/…
 
I see
 
but the generally annoyed tone I got when using restricted-time was more annoying
 
i see
What is the maximum time
 
6:52 PM
I set 10 seconds and I see time limits all over ppcg time limits
but people really got upset about it so I removed it
 
Okay
Then you should remove the tag
 
take this question for example codegolf.stackexchange.com/questions/129716/…
that has lots of upvotes
 
@Anush but amortized analysis takes that into consideration. I did not choose it based on big O time analysis.
 
"You code must take less than 30 minutes to run on a standard desktop machine for any input x such that |x| < 2^40."
@Rick the constant are still hidden in the big O notation
 
I see
At least it inspired amazing answers
 
6:54 PM
my personal view is twofold a) there is a fair point that if you want to set strictly objective criteria you can't just say "30 minutes on a standard desktop machine" as that isn't properly specified
 
That's what I would value
 
b) but people are also annoyed because my question was hard to answer at all and they wanted something trivial to golf
I have much more sympathy with a) than b)
 
@Anush Really?
 
what is the really in relation to?
I am bad at guessing context, sorry
 
@Anush I was expressing suprise
 
6:56 PM
but about which part?
 
I expressed surprise at b
 
oh ok.
 
Because most people say that trivial questions are worse
 
the criticism I got was along the lines of "It's stupid to ask a question that is hard to answer at all and then to require it to be golfed. The golfing part is just irrelevant"
 
958
Q: Build a working game of Tetris in Conway's Game of Life

Joe Z.Here is a theoretical question - one that doesn't afford an easy answer in any case, not even the trivial one. In Conway's Game of Life, there exist constructs such as the metapixel which allow the Game of Life to simulate any other Game-of-Life rule system as well. In addition, it is known that...

That was not trivial
And it was supposed to be golfed
 
6:57 PM
yes.. I do find that whenever I posted another question to make my point the reply would be "PPCG was completely different then"
so I couldn't win
 
@Anush I think he means people golf code because they like solving hard problems, on a competitive code golfing site.
 
2 mins ago, by MilkyWay90
@Anush I was expressing suprise
QFT was 2 years ago though
 
@MilkyWay90 sorry just saw it
 
@Rick Okay
 
in any case... confidence restored :)
 
7:02 PM
@MilkyWay90 the question was posted 6 years ago, plus QFT started a long time before it got finished. if anything like it would be posted today i wouldn't think it'd be received well
 
@dzaima Oh, I see
 
how can I find all questions by the deleted user user9206 ?
 
I don't think PPCG is any different than it was when QFT was revealed, and frankly if it is and things like QFT are frowned upon now, I'm not sure I really wanna be part of the community anymore
2
 
QFT is quantum fourier transform to me.. what does it stand for?
 
7:04 PM
Quest for Tetris
 
thanks
 
The question I linked
 
@dzaima are you saying it's a good question and deserves more upvotes?
 
@Rick (S)he doesn't state any opinion on that
 
Men don't get capitals? :)
 
7:05 PM
Just that today it wouldn't be received well, unlike 6 years ago
 
but I see more recent highly upvoted questions like the one I linked
 
@Anush That's just a way of shortening he/she
 
@Skidsdev i mean, you can't really build a site around the concept of "spend a couple years solving one challenge". I feel something like it is a one-time opportunity. It is a good question and an interesting challenge, we just can't have 10 like it posted a day.
 
@dzaima True
 
this seems to be a 2019 invention to exclude slightly less formally restricted-time questions
@MilkyWay90 it was a silly joke about regexps
 
7:06 PM
@Anush Oh
 
@MilkyWay90 Now I'm conflicted.
 
@Anush That question seemed okay in terms of difficulty, but You may not use any libraries or builtin functions that perform modulo arithmetic (or this inverse operation). This means you can't even do a % b without implementing % yourself. You can use all other non-modulo arithmetic builtin functions however. seems too unclear
2 mins ago, by MilkyWay90
Just that today it wouldn't be received well, unlike 6 years ago
 
@MilkyWay90 fine but it's the time limit which was ok then
but somehow now would be taboo
it wasn't posted that long ago
 
We've developed a lot in terms of rules since 2017
 
hmm
 
7:09 PM
Although I may be wrong, I've only been here since late 2018
 
@Anush i can't really put my finger on why, but that question (except that stupid restriction of no modulo built-ins (would a C compiler optimizing to modulo anyway count?)) by some reason feels more "possible"/"approachable" than your powers question
 
@dzaima but that's a comment about hardness isn't it?
I mean the power tower question is hard
although now cracked with awesome answers
 
@dzaima C optimizes C++ does not for %
 
@dzaima The power tower I guess requires more math
 
I mean if you want hard, check out codegolf.stackexchange.com/questions/77051/…
again with a completely awesome answer
 
7:11 PM
I think it was poorly received because you posted another question vaguely similar right before that question
 
@MilkyWay90 yes that is right
but I worry that @dzaima hates it anyway because of the restricted-time restriction
and I feel he/she isn't the only one
 
Hmmm
Maybe
 
I do get that my type of question is not the same as the "show off how amazing your esolang is" question
but surely ppcg is big enough for that
 
But Magic Octopus Urn mentioned your previous question
 
@Anush now that i actually read the question, i see that it's about modular inversing which is a widely known, solved problem, whereas your question required quite some new proofs (which are not code, on a coding site) to even have an algorithm
 
7:13 PM
@dzaima yes!
is that bad?
 
@dzaima I don't see how that is bad
 
@MilkyWay90 why are guys to talking about dzaima in the third person when they are in the room?
 
@Rick :)
that's funny because they is also third person (except not here I know)
 
@MilkyWay90 alone it isn't bad, but it'd be more suited as a question somewhere else as it's really asking for proofs/math while it shows itself here as a question
 
@dzaima but this is the programming competition website on the internet
 
7:16 PM
@Anush I wanted to be consistent with the flow of the conversation
 
my view is that it should allow PP as well as CG
 
@Anush there's math.SE, where the same question was posted, and i think it's way more suited there (or cs.SE)
 
But again, I've here only a few months
 
@dzaima nooo!
 
@dzaima That's true
 
7:17 PM
If it's ok in the IOI or Kattis it should be ok here
imho
some of the kattis questions are reeeallly hard
 
@Anush though i don't know what exactly IOI is, I'm pretty sure that's not a true statement for PPCG at all. PPCG often it quite opposite of olympiads
2
 
@Anush just upvoted
 
@Rick thanks
 
@Anush you know, the PP part of the name might be going away soon (as always, in 6-8 weeks)
 
@dzaima I do respect your view on this but I hope you don't mind I don't agree
@dzaima I read the suggestions. Very few seem to be designed to remove non code-golf questions
 
7:20 PM
@Anush S/he means it literally
Whoops
Accidentally replaced literally with figuratively
 
you literally replaced figuratively with literally
 
We've decided our site name is just Code Golf now
 
yes I read the answers
 
That's why dzaima means, I think
 
the most upvoted answer is Code Golf
& coding challenges
the second most upvoted is Programming Puzzles &
Code Golf
etc.
 
7:24 PM
We've already decided that it's Code Golf with a subtitle
 
ok
I am not sure it relates much to this conversation
 
I feel Golfing code should be more closely linked to math
 
@Rick I'm okay with both
math ones and other ones
 
@Rick i don't feel that way, but i might be biased because i mainly golf questions
 
7:27 PM
I wonder what the most beautiful image you can make with x ascii characters is, for each x from 1 up to 100 say
I want to ask a code-golf pinball game question ... but I won't :)
 
@Anush To your big power question, what, in your opinion, does add to the question beyond what the math.SE question can achieve?
 
@dzaima it requires you to think of a method which solves the exact problem in such a way that it can be coded efficiently.
rather than a general mathematical solution
this is exactly what the new answers seem to have achieved
unlike the older one
 
@Anush Why not?
 
@MilkyWay90 I fear downvotes now for anything that isn't a traditional code-golf-this-in-an-esolang question
 
my argument is if your code is going to be short the model the code is based on needs to be proportionality compact.
 
7:33 PM
@Anush doesn't that effectively come down to "try to find as simple as possible of an algorithm"? simple tends to very often be proportional to golfiness
 
@Anush Who cares about downvotes when you have creative answer on that question?
1
Q: Compressing a DNA-like string

AwokeKnowingmy app is multiple choice test, where for each question the response results in a 4 letter string "eg" GTAC or ATGC or CATG, etc. There are always just 24 questions. so the final result is something like GTAC CATG TACG GACT GTAC CATG TACG GACT GTAC CATG TACG GACT GTAC CATG TACG GACT GTAC CATG TA...

Reopen this
It has a winning criterion now
 
@MilkyWay90 :)
@dzaima in a sense and I see algorithms and coding as very similar
 
@Anush I think DS might be a more appropriate choice
 
@Anush from that description i can't tell much, but if you use the sandbox, downvotes shouldn't be a problem (or at least they would be predictable)
 
@Anush Ummm
 
7:38 PM
@Anush but both have different sites - coding & golfing here, algorithms in math.SE/whatever
 
2
Q: pwaS eht tirsf dna tasl setterl fo hace dorw

Comrade SparklePonyOr, "Swap the first and last letters of each word" Your challenge is to, given a string of alphabetical ASCII characters as well as one other character to use as a delimiter (to separate each word), swap the first and last letters of each word. If there is a one-character word, leave it alone. ...

 
@dzaima I definitely don't agree with that!
there is even a tag
coding and algorithms really have a very large overlap
 
@Anush okay that blends the two together in a nice way, but what that precisely means, i have no idea ¯\_(ツ)_/¯
 
@Anush that's my favorite tag. However, what's fast might not always be optimal. I think they should change it to optimal-algorithm.
 
@dzaima ok. I feel we are coming from completely different worlds
but that's ok
@Rick you can't know if your algorithm is optimal
 
7:44 PM
@Anush if you overlap them, it's okay. But if you join the two together with glue by slapping on a theoretically interesting & difficult question, is where i start having problems.
 
I have to go. but thanks for chatting
I think we may have to agree to disagree
 
@dzaima I might like that better restricted-complexity
 
@Rick what I like in is that it isn't necessarily the fastest, that there's often room for improvement not necessarily by getting better theoretical speed
 
@dzaima I see
 
7:50 PM
@dzaima now that's problematic, what happens if a new CPU comes out that makes an expensive operation cheap. That's shaky ground to be on.
 
@Rick then time for new algorithms!
 
This is the first argument I've witnessed without devolving into "NO U" chains
 
@dzaima this statement makes me conflicted. I agree and disagree.
 
@dzaima though, with that logic, big powers also has that with golfing.. ok i have no idea why my thinking about ppcg is as it is, but it is
 
@dzaima Oh
 
7:57 PM
Everyone now hates on linked lists because CPUs now use a lot of local caching. However, I don't like that bandwagon mentality. It's not productive and lopsided, it's problematic.
 
8:08 PM
@MilkyWay90 Nothing else from me on the Sandbox post. Let it cook a couple more days and see what others think.
 
8:58 PM
0
Q: Save my secrets!

QuinnI have so many secrets and nowhere to keep them! The goal is simple: write a program that lets you save a string and have it be protected by a password. The program will take an alias, password and (optional) secret as parameters. If this is the first time the program is called with a given al...

 
@AdmBorkBork Okay
0
A: Sandbox for Proposed Challenges

MilkyWay90code-golfnumber To exponential digit growth and beyond! Challenge Given a base (which is less than 10) and a term, your mission (should you choose to accept it) is to: Make a number which will be the base added by 1. We will call this "the current number" Do the following the term number of ...

Repost
 
9:15 PM
yay i wasted today adding tab management to processing
 
What's the music starting at 0:24?
 
 
3 hours later…
11:53 PM
If anyone here has node: does npm i -g vsl-lang/vsl do anything for you or does it just hang?
 

« first day (3029 days earlier)      last day (1814 days later) »