« first day (1192 days earlier)      last day (3645 days later) » 

12:10 AM
I want to do a challenge that would be judged based on the best compression of a file. This would be difficult for me to test.
Any ideas
?
 
12:21 AM
@TheDoctor I don't quite understand. Is the challenge to compress a file? Or is compression only used for scoring? In which case, is it the source or the output that's compressed? And where is the difficulty?
 
The challenge is to compress a file
 
Then where is the difficulty in testing?
 
Downloading all the compilers/interpreters
 
Oh so you're just saying you don't trust the compression sizes the participants are claiming? Or do you not want to give out the actual benchmark file, but just some example input?
In any case, I'm sure you can ask for help here to test the obscurer languages.
 
I will give out a benchmark file generator (the files it makes are ~45 MB)
 
12:30 AM
Oh but because they are random you want to test the submissions yourself against x generated files for scoring? (so that people can't keep regenerating until they find a lucky input for their algorithm)
(btw, we recently had this with the monopoly challenge: even if the decompressor is irrelevant for your scoring system, require one anyway, as proof that the compression isn't lossy)
 
That's what i was planning. I will also disallow solutions that would never really work (like hashing)
 
Re testing, I guess you can ask for compiled binaries if you don't want to install a particular compiler/interpreter. Otherwise ask for help here.
 
alternatively, let your generator take a seed. and then the scoring will be determined from the total compressed size of the files generated from x fixed seeds. this way everyone could do the testing themselves. (of course, that would still require trusting people that the score they claim to have is actually based on on those compressed files)
of course that would add a deadline to the challenge (the day on which you publish those fixed seeds)
 
That's why i would have to it myself...
 
12:45 AM
I see, fair enough. Well I suppose that's just the same dilemma as for everyone posting a KotH challenge.
 
You kinda have to test those yourself...
 
 
1 hour later…
1:56 AM
Hi :D
Working on an event-driven expression-based Lispish language w/ lambda calculus
might be useful for golfing, might not
 
hi. new here?
 
New to the chat, not the site though
augh, Python 2 beats my language at determining whether a number is odd or even by 3 characters
probably because of how stupid operator calls are atm
 
your language is?
 
@TheDoctor Maybe gonna call it Ena, it's quite lisp-like
wanna see some code? :3
 
sure...
how about lithp?
 
2:07 AM
@TheDoctor :P one sec, pasting somewhere
4:10 am is probably not THE best time to design languages...
4
 
I see. I prefer Python 2! :)
 
Heh. I've always wanted to design my own language meant to do something useful, and then abuse the heck out of it by golfing it down to ugliness :D
 
@Doorknob xD Heeeeyy
@Doorknob The weird part is, I haven't even gotten to that second part yet
 
:P
-3
A: Bring out the inner llama of a sentence

Billy JoelFirst of all, what the heck is the inner llama of a sentence??? This makes absolutely no sense whatsoever. Is it like you're trying to persuade the sentence to come out of the closet or something??? It sure sounds like it. That made no sense, but neither did the question, so who cares... Billy Jo...

^ lol, flaggy plz
(6 offensive/abusive flags = insta-delete)
 
I used to like Billy Joel :P
 
2:14 AM
flagged :)
 
woops. i did non-answer
 
@TheDoctor ...me too D:
hehe
 
yo
hey
[beatboxing]
 
just wrote code for the golf doorknob linked to :P
Trying to write some documentation so I can remember Ena (???) tomorrow :P
always been bad at writing docs
 
2:54 AM
I haven't seen @mniip in a while
 
 
1 hour later…
4:05 AM
@Doorknob Delete voted. I already flagged as not-an-answer.
 
4:20 AM
Is there a reason why the map size for this question:
154
Q: Survival Game - Create Your Wolf

RusherThe Board The Board is a two dimensional array of cells. Cells are populated by Animals. Every day, all Animals on the Board simultaneously make one move. If two or more Animals move to the same cell, they fight until one remains. The possible moves and attacks are as follows: Moves - { Move.U...

is sqrt(n+3)*20 and not sqrt((n+3)*20)?
The second would add 20 squares per entry.
The first adds 400. Oh.
Duh, it is easy to move the *20 in by multiplying by 20, so the map size given is sqrt((n+3)*400)
 
 
3 hours later…
7:45 AM
I'm working on a Revenge of the 5th KotH which will probably only be ready be the 6th. :-(.
Just because the testing program is really hard to write. I have 350+ lines of code.
 
 
1 hour later…
9:04 AM
@Quincunx Sounds cool :D
 
9:55 AM
I've started on project Euler using J, and I've gotta say: I'm really starting to like it.
 
Project Euler or J?
 
J
I already started on Euler with Python, but J is just more fun to program in
 
I use python for Euler, but I'm a terrible programmer so I do a lot of silly things
There's one problem where you have to find all equations of the form "x * y = n" where the digits of x, y, and n together have every number from 1 to 9 exactly once.
I generated every possible sequence of digits from 1 - 9, stuck a '*' and '==' in every possible spot, and used eval to see if the resulting equation was true.
 
Sounds familiar.
There was a (now deleted) question here, that asked "how can you use the digits 1 to 9, in order, with operators between them be used to make 100?", and that was my exact sollution
Even though I did use eval("blah")==100, because that's more efficient ;)
Also: while I don't think that the bible was handed down by God to man, itertools most certainly was.
 
10:23 AM
it'd be a simple itertools.permutations loop o:
 
 
1 hour later…
11:37 AM
I was thinking of posing a series of themed KoTH challenges however I am not sure how an idea would be received.
The overarching theme was to be be Fighter Jet Fight, with the individual challenges to be something similar to:
1. AI to engage/avoid fighter jets (The AI with the most kills over 10 rounds wins)
2. target acquisition for a Infra Red missile (the code that successfully acquires the target over progressivily more cluttered scenes wins)
3. missile tracking and ai for missile movement to target. (code that scores the most points where points are given to damage to target)
 
11:59 AM
@Moogie Everybody loves a good KoTH :D You might want to integrate your individual challenges into a single entry, it's more satisfying if you can set up an entire fighter jet
 
@Trimsty I was thinking the same thing, but it would be pretty overwhelming to have a single challenge involving all the individual challenges.
Perhaps doing it incrementally, with the current "best" solution for a individual challenge being used to power the game engine for the next challenge component.
i.e. the best target acquisition solution is used as input to the missile tracking competition
hmm... I am thinking I am starting to like the idea of the component based challenge... I am envisioning that the challenge starts with some very rudimentary solutions for the individual components that the competitors are able to then propose replacement solutions. The scoring would have to be done slightly different though.
Perhaps scoring based on a exponential of the the improvement to the component performance combined with the improvement with the scenario outcome (aka target hit score)
 
12:56 PM
I think this needs to be reopened
1
Q: Unique Base 64 Incrementor With Substituion

6ft DanThe goal is to write a script that will use your (Base 64) numbering system and never repeat any character within the string while incrementing. When you rollover a big number that jumps to the next place value you need to substitute it with the next unique in place... example (in Base 10) 0985,...

 
@Moogie I really like your idea of using the previous best submission to fuel the next one. You could build something pretty neat with the power of 20 people helping you for free :)
 
Next time anyone sees a poorly written post, whether is be just hard to read or flat out a bad challenge, point them here.
 
I can't click on that
 
1
A: Template For Challenges

QuincunxYou can follow this format: Introduction Briefly describe the challenge. Provide a short background for your challenge. Briefly answer the following questions for your readers. Why is this challenge interesting? Did you create the challenge? Give credit and provide links to your sources. ...

It is community wiki and it's kind of new, so edit it if you feel the wording is off.
 
1:06 PM
Fun. Ok. @Door see above
 
Noted; perhaps I'll edit it a bit after school today
... which is exactly where I'm going now. Bye all
 
And the delvotes
 
Since when did fastest code become a poor choice for win condition?
This guy originally had fastest-code as his winning criterion, and then @m.buettner challenged it, he removed it, and now the challenge sucks. codegolf.stackexchange.com/q/26551/18487
 
Difficult to be objective. It requires the asker to test each answer individually.
Darn
 
I just searched for [fastest-code] and there are tons of good recent challenges using that tag as their only winning criterion
Are we torching that one along with code trolling or what?
 
1:16 PM
I think there were other problems with that post in particular.
 
yep
 
For instance, Revision 4 makes it look like he's just looking for a better big O, and not actually concerned with fastest runtime. It looks more like an assignment than something meant to be a challenge.
Not to say it couldn't have made for a good challenge, but unless it's edited into shape, I couldn't vote to reopen it.
 
Ah, I didn't see the revisions. I just wanted to see if that tag was being torched as well.
Glad to hear that it isn't. I think it captures a side of coding that is pretty important
 
Yea, I don't think it's going anywhere.
 
I like it better than code-trolling, and it's more objective than popcon
 
1:22 PM
@Rusher you should see from my comments that I didn't close-vote and challenge the fact that it's fastest-code, but that it wasn't clear what the actual winning criterion is. Tagging something fastest-code alone is not sufficient, because there are a bunch of ways to interpret that. (Algorithm complexity, actual runtime, etc... and there multiple ways to do each of those as well)
 
So no, I do like fastest-code challenge but that one in particular wasn't well put in my opinion. In addition it really seemed like the OP just wanted to get help with his code but got rejected on SO.
 
@m.buettner Well, that's actually wrong. From the tag wiki - This tag indicates that the program with the fastest runtime speed is the winner of the challenge.
 
@Rusher in that case, what's the benchmark input to measure runtime? And is the OP willing to test all submissions on his own machine?
 
It has been customary so far to allow challenges like "fastest runtime on any reasonable desktop". Why is that a problem now?
 
1:25 PM
Okay, I wasn't aware of that. That still leaves the problem of a fixed benchmark.
 
Believe me, I challenged Doorknob when he wrote that exact phrase on his challenge, but I got negative feedback, and I've since accepted that posting a full set of specs is probably overkill
 
The other problem was that the OP didn't actually seem interested in making this a general challenge though. I'm not sure if you can see deleted answers, but I provided on in Mathematica, and the comment I got was "Thanks for this. Would you mind additionally demonstrating how the same could be done in an object oriented language, like C# or Java? I've been struggling to get this done for a website code written in .NET where search relies on this technique."
Hence, his post was off-topic - not even because it was an underspecified challenge (and definitely not because it was a fastest-code challenge) - but because the OP really wanted answers he would rather get on codereview than codegolf.
 
Yea, I agree that it had other problems. I just thought that after editing in his own solution to the problem and removing the only winning criterion, it got worse. Don't want to confuse new people into thinking that "If you are new, you had better stick to the code-golf tag or we'll torch you."
 
Yeah, I agree. But I tried to make it clear in my comment that the problem was just that the winning criterion was underspecified so I made some suggestions for how to actually measure it. Sorry if that wasn't clear.
 
Kind of along those lines - what does this show when you click it? repl.it/SSO
Does it show my stopwatch?
 
1:30 PM
Yes
(why am I clicking on cryptic links posted by strangers? :D)
 
Sweet! If I were going to run a fastest-code challenge, I would use repl.it as the test machine.
Well, mostly because repl.it is a well-known online interpreter for a bunch of different languages
Kind of like js fiddle
Anyway, that stopwatch accepts a function, a number of iterations, and args to that function (in a list)
 
Well, now I know it, too ;) ... thanks for introducing me! :)
 
It runs it that many iterations and returns and average clock time
 
you know, there is also ideone. Not a REPL, but the number of supported languages is much larger.
 
I think ideone does python 3
 
1:32 PM
ideone has... other problems if you're trying to test for speed
 
repl.it is python 2
 
ideone does both
 
there's a strict timeout, for instance
 
Bash
Pascal (fpc)
C
Pascal (gpc)
C#
Perl
C++ 4.8.1
PHP
C++11
Python
Haskell
Python 3
Java
Ruby
Java7
SQL
Objective-C
VB.NET
Ada
COBOL 85
Intercal
Perl 6
Assembler
Common Lisp (clisp)
JavaScript (rhino)
Pike
Assembler
D (dmd)
JavaScript (spidermonkey)
Prolog (gnu)
AWK (gawk)
Erlang
Lua
Prolog (swi)
AWK (mawk)
F#
Nemerle
R
bc
Factor
Nice
Scala
Brainf**k
Falcon
Nimrod
Scheme (guile)
C++ 4.3.2
Forth
Node.js
Smalltalk
C99 strict
Fortran
Ocaml
Tcl
CLIPS
Go
Octave
Text
Clojure
Groovy
Oz
Unlambda
COBOL
oops, sorry
 
repl.it has a timeout as well, somewhere around 15 seconds i think
 
1:34 PM
Ah, didn't know that.
 
Anyway, if everyone wrote a stopwatch for their favorite language and added it to the wiki for fastest code, it could be made a lot less objective
 
The other problem I see (which could be an issue anywhere) is input size. I know on ideone the total size (code, input, output) is limited to 64kB. Could be a problem if your challenge is something like "crunch these million numbers", etc.
 
Oh man I am that guy now :(
 
Well... right, but then the function to generate and crunch has to work in the time limit.
 
I suppose it's a viable option for some challenges though, where neither code nor input nor runtime should be too large for submissions anyway.
 
1:38 PM
I can't stand when another programmer says "Well, why don't you just do it this way? It's obviously so much better." So slap me if I do it to you.
Unless your talking about implementation details or performance... then it's fine
 
Yea, for some maybe. I'd think most of the best ones would allow more than 15 seconds, even if the winning entries don't take that long.
 
If you restrict it to Python, then you can swap clock time for CPU time and just run all the challenges on your own computer. Would that be fair, or are different CPUs better at "specific" tasks?
I think I have a stopwatch somewhere for cpu time... but not with me
 
There are benchmarking tools available to do all this, though. Just a matter of taking the time to test it right.
(don't ask me for the name of one, just trust that they exist)
 
@m.buettner I can't bring myself to read the middle of your book on meta. I got the beginning and the end, though
Needs headers or something
Nvm, I made it through. The middle paragraph was the most important one anyway. Especially "They all take a bad question from SO (or make one up), and post it with the code-trolling tag as a one-line"
You know you are a bad troll when you are making up a bad question from SO and posting it here.
 
Eh, I prefer a made-up one over one that's copied over with a link, essentially saying "look what this idiot posted".
That's just.... rude.
3
 
1:51 PM
Sorry, I just had a revelation that a "bad question" is identical to a "bad question that I made up"
 
Also, I find it funny that people didn't like the self-censoring of "balls" in that post, but freak out if someone says Brainfuck.
 
I have a policy of just posting words spelled correctly. If there is a censor, it will kick in. If you misspell things and the censor misses it, then you can get banned
 
like "Brainfcuk"?
Also, @Geobits, there's a difference in the average maturity level of people who actually read meta compared to people on the main site.
 
@Rusher thanks for persevering ;) .... I'm sorry it became such a tome, but I thought I'd rather present a solid (if lengthy) argument
 
@JoeZ. Maybe so, but I just thought it funny. Not a big deal or anything.
 
1:55 PM
Maturity, not age
 
Yeah, true.
I don't think anybody who complained about "b***s" would complain about "Brainfuck".
 
@m.buettner You know if you click the down arrow to the left of the message you are responding to, then it will automatically put my name like @Rusher and it will also highlight the message you are responding to
 
Especially me, I posted that "JSFuck" question a while back.
 
Otherwise, it uses most recent
 
@Rusher Yes, I'm aware of that, I'm just still in the habit of writing comments, where I just type the first letter of the name and hit tab. ;) (also clicking the respond icon on the right is quicker than fishing out the option from the arrow menu)
 
1:57 PM
What does b***s stand for anyway?
 
bands
banks
barfs
 
boxes
 
baits
 
@m.buettner Oh ok. Someone had to tell me before I noticed lol
 
bares
 
1:58 PM
@Geobits you're too focused on that a
 
Working my way through
 
bolas
 
@Rusher still, I need a mouse for that... really annoying :D
 
burns
 
bunks
 
1:59 PM
bouts
beats
 
bells
bulls
bills
 
bings ?
 
nah, Microsoft's search engine isn't a generic verb like "google" yet
Who says "go bing it"?
 
................................................................................‌​.................................................................................‌​.................................................................................‌​.................................................................................‌​.................................................................................‌​.................................................................................‌​..............
 
no, like Crosby
 
2:00 PM
who uses bing?
 
who?
Sidney or Bill?
 
I used Bing for a research paper in college
 
At least one building full of employees use it.
No, Bing Crosby, not Bill Cosby
 
ah
and Sidney is Sidney Crosby, the guy who scored the winning goal at Van.2010
I have absolutely no idea who Bing Crosby is
 
Ah, not a fan of the sport
Old-timey singer
 
2:01 PM
I see
Who says "go bing it"? <- you'd sound like you had a "mao bing" in Chinese
"I went and binged a supermarket." just doesn't have the same ring to it
not to mention confusion with another sort of "binged"
 
"Ice cat"??? or am i getting my pinyin wrong?
 
you're getting the intonations wrong
you did mao1 bing1
 
You didn't specify any
 
I was trying to say mao2 bing4
well
 
And you definitely wouldn't bing someone before a date.
 
2:02 PM
yeah, it's a bit hard
LOL
but yeah, "mao bing" intoned that way becomes "problem"
and who would bing their own name to see what comes up
(and "let me bing that for you")
okay, I'll stop
 
I just binged myself
 
Binge - a short period devoted to indulging in an activity to excess, especially drinking alcohol or eating.
 
Yes, I did that, too.
 
That's pronounced totally differently, though.
 
Maybe you could go on a Bing binge?
 
2:05 PM
Perhaps, but meh.
On another note, nobody's making new entries for the Prisoner's Dilemma challenge :<
 
Oh well, better odds for me then :)
 
Isn't that like an ancient challenge? Did we revive it?
 
Yeah, I posted a v2
where everyone plays against each other all at once
 
Oh hey.. I've upvoted it already lol
 
lol
Perhaps write an entry? :P
 
2:07 PM
My Computer Science class would be more interesting if we did stuff like this
 
Pitch it to them. I have a judge program ready for them to use :P
 
@TheDoctor Our class did a heavily modified version of the Wolf challenge I posted
We also wrote minesweeper solvers
 
Fun. mine does boring Java exercises
 
That reminds me, I should probably give my high school CS teacher my judge code so that he can do this challenge as well :P
 
We had those too. Like "Write a shape." "Make a square that IS A shape." "Write an interface that makes shapes <something>able"
 
2:11 PM
We had those too. At the end, though, we were given an assignment to write a program that involved an AI.
 
AI was in special topics (which I never took). They also got to do pathing and other cool stuff
 
I ended up making this: joezeng.com/games/puzzleleague (although in Java)
Well, it was a really simple AI, it would just make random moves unless there were three blocks of the same colour in a column
but it ran in realtime and actually beat me a few times :<
 
2:24 PM
Anyone heard of Golly?
 
Yep. I played with it years ago.
 
It's cool
 
2:37 PM
Oh golly
 
3:29 PM
@JoeZ. Should your Go challenge accept input as arguments or on stdin? One answer doesn't use it, one is using arguments, and one stdin.
 
@Rusher I don't recall seeing any challenges like that, and I can't find any with search. Can you find an example?
 
I think from stdin
@ Geobits
 
4:14 PM
@PeterTaylor Did you delete your own "any reasonable machine" question before asking that?
You wrote a challenge that said something like "must run in under x minutes on any reasonable machine"
You were the one who convinced me that it was objective enough to allow
 
That wasn't a fastest code question.
 
@PeterTaylor My argument stands.
"fastest code on a reasonable desktop" is either too subjective or not. If so, I should find your question and nominate it for close.
 
@Rusher "Fastest code on a reasonable desktop" is not the same as "runs in under a generous amount of time on a reasonable desktop".
2
 
Pigs are not cows, but "runs in under a generous amount of time on a reasonable desktop" is still subjective.
 
A constraint on a code-golf question which aims to avoid naive brute force isn't relevant in a discussion of how fastest-code as a winning criterion should work.
 
4:24 PM
I understand what your goal was, but "My goals were honorable, therefore my spec is objective." is a nonsequiter
Whatever standard we hold to, we should also hold any question that uses performance constraints.
If "fastest on a reasonable desktop" is valid, then "under 2 minutes on a reasonable desktop" should also be valid.
"fastest" and "under 2 minutes" are both objectively measurable. The common denominator here is what a "reasonable desktop" is.
I'm not asking you specifically to admit that your own question is poor. I'm asking for consistency. If your question is valid, then so is "must perform X on a reasonable desktop" as a criteria. If it is valid, we can improve the fastest code tag to include that clause and thereby make it stand alone like the other tags do. Unless the author wants to be more specific.
 
4:42 PM
The difference is that you can only have one "fastest-code" if that's the winning criteria. If it's just a sideline-criteria, it's only for ruling an entry valid/invalid. There's a fine line there, but I'd allow more leeway for things that aren't the primary criteria.
If I said, "should run in under two minutes", I'd probably still allow something that ran in 2.5 on my machine. The author could make a case that on a "reasonable" machine it would run under 2. If I'm trying to pick a winner, there need to be exact results.
 
Sure. If someone tells me that primary criteria have more leeway than secondary criteria, then I'm wrong. I had never heard of that rule before. I was under the impression that you had to meet the spec before you could even be considered as a winner, and I assumed that the spec had to be equally as solid as the scoring mechanism.
 
I'm telling you my opinion. I can't speak for anyone else, but if I'm trying to judge a [fastest-code], it damn well should have a better definition of "fast" than a [code-golf] should.
That said, I think the [fastest-code] wiki is lacking, compared to many of the other criteria.
I think a "What makes a good x..." would be a good addition for all of them.
 
I have the exact opposite opinion. If code-golf mixes in performance criteria, I think it should be as solid as any other performance criteria, whether it be "winning criteria" or "valid criteria"
 
Vote your opinion then; that's what votes are for :)
 
5:29 PM
This discussion is largely a red herring given that the original one was about how you determine which implementation is faster without a benchmark.
But since we're having it, I reason in a similar way to Geobits. If my question has an acceptance criterion that the code should execute the benchmark in less than 2 minutes on a reasonable PC, even if it takes 3 minutes on mine I'll accept a third user's comment that on their reasonable PC it runs in less than two minutes.
But if one program runs faster on my reasonable PC and the other runs faster on someone else's reasonable PC, how can I pick a winner?
So I think the OP has to say "I'll run it on my PC and post timings".
 
In that case, the tag needs to specify that the OP must declare himself the official tester with the official test machine.
I'll edit it now.
FYI, I took what we discussed here and put it on meta: meta.codegolf.stackexchange.com/q/1523/18487
Now that I think about it, I will wait and see what happens to that post before editing the fastest-code tag.
 
6:28 PM
Bresenham's line algorithm... ugh. I just want to be able to iterate through the points along the line, but the wikipedia page looks so complex.
 
 
1 hour later…
7:36 PM
in The 2nd Monitor, 16 secs ago, by Rusher
46 secs ago, by Malachi
in The Nineteenth Byte, 12 secs ago, by Malachi
people in http://chat.stackexchange.com/rooms/8595/the-2nd-monitor are talking about you
 
 
2 hours later…
10:06 PM
-3
Q: Generate all serial combinations of tokens of a given string

SNagGiven a string, This is a test string, first tokenize it, [This,is,a,test,string], then generate all serial combinations of the tokens. The expected output is: [This,is,a,test,string,This is,is a,a test,test string,This is a,is a test,a test string,This is a test,is a test string,This is a test ...

Delvote ^^
 
@TheDoctor for fairness I'd wait until the user has acknowledge that he read my last comment.
 
@TheDoctor I'm honestly tired of the spam-like delvote requests. From Stack Meta, "The system will automatically delete unlocked, unanswered questions that have negative score after 30 days."
Let the system do its job and delete the delete-worthy questions on its own. There must be a reason the community chose 30 days. It is so incredibly specific that it seems silly to force it to occur earlier unless you are terrified that the author might come back and attempt to improve an unsalvageable question.
 
k
I try to run a tight ship, i guess.
 
10:43 PM
0
Q: The easy way to code-golf ascii art!

TheDoctorTask: There are a lot of answers on this site that are arrange into ascii art, like this one. Usually the arrangement is done manually, but wouldn't a program help with that? :) Your program will take 3 inputs: The code, as one single line The number of lines in the pattern (can be omitte...

 

« first day (1192 days earlier)      last day (3645 days later) »