« first day (1369 days earlier)      last day (3480 days later) » 

12:07 AM
@Geobits What do you think of my (impossible?) challenge proposal?
 
It would be better if I could think of a simple way to test it without spamming the crap out of SE and getting banned ;) It would also help to know how they "detect" messages (what can you not do).
 
Well, I was going to post some messages for you to implement, and the winner would be chosen my pop-contest.
Such as "The enemy will attack at dawn"...
 
How to test?
 
You must post your solution (question/answer) containing these words in your answer. You don't need to post on SO.
Also, you must include an introduction and conclusion sentence that does not contain the message (this will help you hide the message better).
 
Ok, so the code doesn't need to post the q/a, just craft the messages.
 
12:13 AM
Yes.
 
That makes a lot more sense. For some reason I thought you had us posting a bunch of nonsense on SO :D
 
Posting isn't the problem, creating a reasonable message is.
No, it is not supposed to be nonsense. The idea is that SO will not close/delete the post because it is quality content.
 
So basically, write code that can craft natural sounding sentences (paragraphs, really) with key words given every n words? That sounds.... hard. Like current-areas-of-study hard.
I meant "nonsense" as in "fake questions", not gibberish.
 
Yeah, write your thesis based on this challenge :)
OK.
 
 
1 hour later…
1:17 AM
0
Q: Answer-dependent questions

YpnypnCalvin's Hobbies recently posted Evolution of Hello World. The challenge is unique in several ways, but the most revolutionary is that the definition of "valid answer" depends on other answers. This inspired several more answer-dependent1 questions. However, as discussed a bit in chat, these que...

 
 
2 hours later…
3:39 AM
Congratulations to PPCG's first and so far only 20k user, @MartinBüttner.
 
 
3 hours later…
6:22 AM
I'm starting to wonder what ways there are to define an operation when writing a compiler/interpreter. I could create an object for each and every operation, but this seems so lengthy. A switch statement on (single char operators) the character could work, then mapping to functions which do the work. But this seems messy. I wonder where I can go to find out a reasonable way to create neat code for a compiler/interpreter.
 
 
2 hours later…
8:00 AM
@Dennis Thank you :)
 
8:31 AM
0
A: Answer-dependent questions

Martin BüttnerI would really like to make this challenge type work. (Full disclosure: partly because I seem to have come up with with a fairly popular challenge idea based on it.) But I do see a lot of problems. Answers on SE have no inherent order, and forcing one onto them just feels odd. In the case of a ...

 
9:28 AM
@BetaDecay The way you've changed the challenge now is exactly what Peter and I were on about in the sandbox. You cannot "exactly reproduce this SVG" without actually outputting that SVG or implementing Bezier curves and stuff. Even then, any PNG or PPM will never be "exactly" like SVG. Which raises the question "how exactly is exactly enough".
In the sandbox you even said "I've changed it to a 'Golf this SVG'." so why change it back?
 
Oh, right I see.
It's back as an SVG.
 
 
2 hours later…
11:37 AM
@COTO "This challenge gets my first downvote ever." Because you overlooked a rule? o.O
 
11:56 AM
@algorithmshark any update for the J tutorial? :)
 
grc
does anyone know if my J answer satisfies the REPL rule?
 
12:15 PM
@MartinBüttner I didn't downvote it. I've explained why the challenge ticks me off in the comments section. But as I also say there, it's my problem, not Calvin's.
 
@COTO Yeah I noticed you ended up not downvoting. Fair enough.
 
 
1 hour later…
1:28 PM
Isn't it funny how you can get the same amount of work done two months in a row, but one month management tells you that you are behind and you feel terrible, and next month management tells you that you are ahead of schedule and you feel amazing?
 
I wanted to make a joke about your amount of work not being a context-insensitive grammar, but I couldn't even come up with a bad one... so instead I'm just reminding everyone of the existence of this wonderful xkcd:
 
I need a voting mechanism that allows voters to add candidates. Can anyone help me out?
And it would be nice if it allowed up to X votes per voter
 
@Rainbolt Wait, do you mean for a challenge, or just for a programming thing?
 
@Rainbolt what do you mean by that?
 
By mechanism I mean an online poll. By add candidates I mean that if the option they want to vote for isn't there, they can add it. And this is for a challenge, but it's a long story
So for example, if I said "What's your favorite language?" and Python wasn't on there, you could add it.
 
1:42 PM
Two options (probably more, but here are two):
1. Add a textbox and if someone can enter Python if it is not in the list, and add Python to the list.
 
Errr.. I'm just looking for a premade online voting platform
 
2. Add a textbox and register how many people enter Python (or something else that's not added), but don't add it to the list.
 
Like this: poll-maker.com
 
ah...
 
But it doesn't do what I want it to do
Oh wow nvm
It does lol.
 
1:45 PM
I don't see an option to add a custom value.
oh wait, now I see
 
Allow users to tick other and type their own answer
But it doesn't seem to allow up to X votes per user
 
Didn't notice the 'options' tab at first.
 
It's just "Allow multiple" or "Allow single"
If you find a way to limit it to 2, let me know
 
if it's for a challenge, make a meta-post?
if it's there I upvote, if it's not there I make a new answer.
 
But then you cannot control the amount of votes.
 
1:48 PM
downvotes are ignored.
oh
I see
overlooked that part
 
Ugh... poll-maker doesn't allow symbols in the answers. A vote for C# just went to nothing, and C++ is showing up as C
Ok, I'll just break it down for you and you guys can maybe come up with a better suggestion
 
Write a survey thing yourself?
 
I want to bundle about ten compilers/interpreters together and call that "the KotH bundle" for those who are lazy
Then I can write a challenge and only allow those ten languages
I don't wanna be the one who chooses ten languages
 
Why do you need restricted votes then?
 
why is it important to limit the number of votes for that?
ninja'd
 
1:52 PM
:P
 
I just didn't golf my question enough
 
Here was my reasoning: I would ask everyone to vote once for their favorite (obscure or otherwise), and once for a well-known language that they are comfortable in.
That way they at least get to participate if their obscure one doesn't win.
 
make two polls?
 
Yes, that^ is also an option.
 
That could work
I can pick the languages for the second poll
 
1:54 PM
Then there should be a way to make C# and C++ work in that poll...
 
C-Sharp
C-Plus-Plus
 
That should work.
 
No symbols
CPlusPlus
 
That's so lame lol
 
1:56 PM
C Plus Plus
 
Pretty soon my entire challenge will be dictated by meta posts
The format for writing the challenge, the format for communicating via cmd args and stdout, the allowed languages, etc.
Once I get to where I can focus on the fun stuff, I'll start writing more challenges
I really want an autogenerated scoreboard, but we're a long way from that
 
With stack snippets that's even more doable now.
Here's what we should do: write an external script, which can be included in a stack snippet, and then be configured via a list of columns, a column to sort by, and a single function, which takes the Markdown of an answer and returns a list of values (i.e. the function parses out the interesting information). That script then generates an ASCII table with all the values returned, sorted by the score.
 
How does it come up with a score?
 
that's what your function does
just stick the score in the "column to sort by"
 
Maybe I'm not seeing the connection between the stack snippet and the bundle of compilers on my machine
 
2:09 PM
there is none
the idea for the script is a bit more generic, not just for KotHs
 
Then this scoreboard isn't "auto" generated. It's manually generated, and auto sorted.
 
that depends on the challenge ;)
but if you've got your automated test runner somewhere, then you can also load the scores from an external resource with that snippet
 
I think that at some point, a forced answer format comes into play
 
yeah sure
how are you going to automate anything without a fixed answer format?
 
Write a "smart" parser?
 
2:13 PM
The forced answer format could be as simple as "put the score, but not other numbers, in a header". Sure some people will still post it wrong, but it allows it to be a bit more flexible.
The biggest problem then would be python 3 :P
 
Oh, we're having two different conversations. One is "parse the scores and generate a scoreboard" and the other is "scrape the answers for code, compile it, run it, score it, and generate a scoreboard"
 
Oh, you're trying to automate the entire process? Yea, that's tough :)
Not least because people tend to put things other than code in code blocks (local results, sample output, etc).
 
well you could just go with the first code block
 
I'm very tempted to say that if you can't follow a well laid out format, then you just won't participate. On the other hand, this probably adds a huge chunk of text to the question that could otherwise be used to write a clear challenge.
First code block seems simply put.
 
You'd also have to exclude posts that are not a entry, but a "helper", like the frameworks, visualizers, etc that people like to post.
Probably easiest to just ban it in the rules.
 
2:22 PM
Aww but I like those things
They add flavor
And they usually get community wikid anyway
 
You could allow them but tell people to label them a certain way to parse them out. I said easiest is to ban them ;)
 
Assuming I have to parse out the user, it wouldn't be a huge leap to parse out the community wiki bit. They are close together in the HTML
I can exclude all CW posts
 
Makes sense. That way if you want to add "non-scoring" example bots you can just CW.
 
Haha I get all the reps for my example bots.
But they also participate, so that's ok
 
Yea, they normally score pretty well for an added rep boost :D
FGITW and all that.
 
2:26 PM
@Rainbolt I think we had a meta discussion somewhere that it doesn't belong in answer if it's not an answer (even if it's CW)
 
To be fair, I usually bounty the challenge. And maybe it even lessens the FGITW effect because it all gets channeled into a bounty.
But I probably get rep for that too with the added visibility
 
@Geobits I think the main rep boost of providing your own answer is the higher chance to get into the HNQ.
 
That too. For a real boost, add 4 answers :P
It goes straight into HNQ then.
 
I have to focus on stuff for a few hours. BBL
 
The code golf tag wiki says that a code golf should at a minimum specify whether you should do a whole program or function. If this is not mentioned in the question (which is otherwise fine), is there a default? I assume not, but I'd rather not have to comment on the many(?) posts that don't specify before every answer I give.
 
2:32 PM
I don't think there is a default.
 
For example, Character Cross doesn't specify, but some of the answers could be significantly shortened if a function is allowed.
 
btw, did you guys see this dodgy question on Arqade? gaming.stackexchange.com/q/189843/86987 :D
<-- needs food
later
 
See you!
 
@Geobits They should be posted as links in comments, not as answers, anyway.
 
I know. I agree with that, but people do it anyway. It's hard to delete them once they've been upvoted though (which they invariably are), so you'd still need to parse them out somehow.
 
2:45 PM
Flag as "not an answer" and get a mod to convert them to a comment.
 
Right, I get that, but if you're running a script to scrape entries from the answers, it won't know to wait until the flag has been acted upon.
 
Yes, I was addressing the "hard to delete them" objection.
 
Yea, I didn't really mean "hard", just not "quick" I guess. I've had flags linger for days sometimes.
 
3:14 PM
@PeterTaylor so... now that you're almost done... what's the next retagging project? ;)
 
Good question. Do you think I should retag as ? ;)
 
Lol :D
 
I present the 25,000th answer: codegolf.stackexchange.com/a/40627/8478
@PeterTaylor well, there was the discussion
I also think is the mother of all meta tags, but I wanted to start a more general discussion about "our" definition of meta tag first (the usual definition doesn't make sense here, because every non-challenge tag isn't valid on its own)
btw @PeterTaylor in golfscript can you convert between codepoints and characters/strings?
 
6
A: Tips for golfing in GolfScript

Peter TaylorTo address a specific subquestion: How to best convert a string into an array of ASCII codes and back? Which operations do this as a side effect? What's the best way to dump the characters in a string onto the stack (like ~ does for arrays)? For those who don't understand the problem, GolfS...

 
3:30 PM
thanks
okay, I don't think that's still possible for the 95 quotes challenge
 
3:43 PM
I think it is if you're hacky enough.
 
well I was looking for a way to construct strings which I can eval, to get around having to use block braces and stuff... but if I need '' or "" to turn a list of ASCII codes into a string, it seems to be a bit of a catch-22
 
That's the easy way to do it. I think you've still got :, so you can have as many empty strings as you need. The biggest problem is the lack of + or *. If you can do something useful with strings with no duplicate characters then ] and | will do the job.
 
what can I do with :?
 
In fact, you can get around the uniqueness problem by using codepoints which add increasing multiples of 256. The displayed output is still ok
Assign the empty string from stdin to a suitable unused character.
 
ah, I see
let me try
 
3:53 PM
The lack of 3 and 0 might be a problem, but online demo
Actually, just add multiples of 256 until you have no 3 or 0.
 
huh, why does ^ work there?
ah, that's why you were talking about unique characters
 
4:04 PM
I do believe that mono for mac does not adhere to the C# spec for decimals
 
done
@PeterTaylor thank you!
what's the most evil thing one can forbid? :D
 
just had a program tight looping on a friend's machine because I forgot I was using decimals (thought I was using ints), and iteratively dividing by 2 was never reaching zero, which the spec says it should. instead, mono mac was getting stuck on 0.0000000000000000000000000001
whereas the spec says it should round to 0 on division
this could I suppose be regional settings, but I'm pretty sure the spec is the spec and the spec says it should be bankers' rounding
 
4:33 PM
ugh, the new reject-edit reasons are terrible for PPCG
 
@MartinBüttner You've used a 3 in your code...
 
@BetaDecay you're right. fixing in a second
oh there are even two of them
 
Yeah... I noticed the last one because it was at the end of the line on the online interpreter
 
fixed
you need a bigger resolution ;)
it only uses two thirds of my screen on the smaller monitor :D
 
Haha :)
 
4:47 PM
oh lol, mobile
 
It fits quite nicely on a mobile screen
Hey, how can the Community account approve edits? I thought it was autonomous...
 
where did that happen?
 
@BetaDecay The OP edited it and marked the suggested edit as helpful.
Then Community approves it.
 
Ohh right
 
4:53 PM
45
A: Why does the Community ♦ user approve and reject edits?

EmrakulWhen a user suggests an edit, their edit is put in the Suggested Edits review queue. The edit will leave the queue under three conditions: If two users (three users on Stack Overflow only) vote to approve or reject the edit If the original poster or a diamond moderator approves or rejects the ...

 
5:06 PM
@PeterTaylor your code seems to fail if you get to an empty string
try 9 generations for 00001 (and the same productions)
but yay, thanks to your submission I looked into mine again and found a way to deal with this in two bytes less
*four bytes shorter
(plus another one for not using strings)
 
@MartinBüttner I missed that one. This is why questions should have proper test suites.
 
agreed
 
6:08 PM
Does anyone here have any experience with CyberCoders? I'm going to be on the market soon and they have a ton of listings on major job sites. I've seen mixed reviews online, so just curious.
 
never heard of that
 
I think it's US-based only, so that doesn't surprise me much.
 
7:02 PM
Because that would be stupid...?
 
lol
I realised once again today that Arqade is still too small a community to be useful for games with less than a few million players
 
I've seen a few times where a blatantly wrong answer is accepted/upvoted simply because it's the only one (I assume). I have much better luck with Google for game related info.
 
well I figured, the situation won't improve if I don't ask questions about semi-popular games there, so I gave it a shot anyway
I also did find a question on a related game that had a wrong accepted answer :D
 
True, never hurts to try. Who knows, maybe you'll get an answer at some point.
 
also, outside of Japan the game was only released last week...
 
7:14 PM
Half the questions there could probably be closed as "just google it", to be honest. That's true of certain other stacks, also, but the wealth of game information out there is astounding, and any "big" games have multiple huge sites/wikis devoted to them.
 
This is the weirdest thing I've ever seen on SO: stackoverflow.com/q/26638703/3224483
I upvoted because it made me laugh to hard when I found it
 
amazing :D
 
I know that makes me a bad person
 
I knew that was you :P
 
aaaaand it's gone
 
7:17 PM
@Rainbolt although I think that award would go to this thing I saw earlier today: stackoverflow.com/questions/5508110/…
5
 
@Martin I love that one :D
 
lol, when I read it this morning I wanted to check whether it was posted on 1 Apr... I forgot to, and just did... of course it was... :D
 
I love that it's been closed 5 times, and locked and protected several times each.
SO is quite schizo :D
 
that's what you get with community moderation :D
it's basically like code trolling, just funnier
 
8:03 PM
@Geobits Good point about the scoring.
Although it's not like it incentivises playing early instead of late.
 
If I can answer an early, simple pattern in 4 bytes, that's 0.1333..., which would take several entries to match if I answered once the average complexity was over 15 or so characters.
 
yes, I agree with that
but obviously your score would be even better if you answered early and late
but still...
I was trying to come up with a scoring that incentivises late and short answers
 
That's true, but the first domino had 30 answers before I even saw it, so "play early" isn't so easy ;)
 
but everything I came up with either gave the victory to the last answer or implied that there is no point in submitting an answer that's longer than its predecessor
 
I agree there's probably nothing perfect, just pointing it out.
 
8:07 PM
well I'd love to come up with something better, so any feedback is very welcome ^^
 
I'll have to think about it. If I come up with something else I'll let you know.
 
thanks
 
@geobits I tend to assume that specs want a program when they don't specify between program or function
I imagine that the asker wants code that does the thing, not a definition of a function that would do the thing if called
 
I tend to assume that (and have answered that way in the past), but I'm tempted to just start posting functions. It really is the very first thing in the wiki's "You must include this" list.
 
either way, i'd be in favor of a default
it would be awkward if people started posted python functions to the Character Cross question that are shorter than current answers but only because they are functions
 
8:15 PM
I almost did that with my Java answer to it, but decided it beat the other one by enough already that it would be a bit unsporting to make it that much shorter ;)
 
grrrrr, 60 second answer limit...
 
Never have I felt as much justification for code trolling as when seeing this (now-deleted) answer: codegolf.stackexchange.com/a/40592/20260
 
That's beautiful.
 
for those that can't see deleted answers, it reads: "Can anyone please provide this in C++ too? :( P.S : I'm the one who asked this 'homework' question at Stackoverflow"
 
1
Q: Default for Code Golf: Program, Function or Snippet?

Martin BüttnerOne of the things many (new?) users forget to specify in code-golf challenges is whether the answers should provide a full program or a function (or even just a snippet). For most such things, we have defaults in the tag wiki (like scoring by bytes), but not for this. The conclusion in chat was ...

 
8:48 PM
@Doorknob lol, thanks for the edit
 
;)
 
9:41 PM
What do you think of this challenge?
14
Q: ASCII's 95 Characters...95 Movie Quotes

Calvin's HobbiesIn 2005 the American Film Institute produced AFI's 100 Years...100 Movie Quotes, a list of the best quotes in American cinema. Here they are listed exactly as they should be used in this challenge: 1. "Frankly, my dear, I don't give a damn." - Rhett Butler 2. "I'm gonna make him an offer he can'...

I can't help because I don't know these ridiculous GS and BF languages :(
 
yeah, the challenge has already reached a really difficult stage
 
I should start another one stating that the languages used must be provably used in major production projects.
This would stop them (Or would it? What is the definition of Major?)
 
exactly. and excluding languages just because you don't know them is very much frowned upon and would probably get you a lot of downvotes.
 
You mean just because I don't know them?
No, I only know C++ and C# to any reasonable extent. Java and Python a little, so that leaves Ruby (and ?) that I don't know.
 
that's exactly what I've written :P
 
9:47 PM
Yeah, it isn't fair. I don't have time to learn GS because I need to do my school work and learn programming so I can get a job in about 3-4 years.
 
but no one requires you to know every obscure language out there
that doesn't mean you should ruin the fun for others who know and like to use those languages ;)
 
Exactly.
But then I am at a disadvantage or outright banned because I don't know them.
Catch-22.
I want to become the president some day and confuse everybody with catch-22's until I can do what I please :)
Just kidding.
 
I think the challenge you linked to has reached a stage where everyone needs to look into new languages to provide an answer... and even if not, people have to look for very obscure features of their language.
 
Oh.
Whitespace might be able to do it with tabs and returns, but the space is banned, so maybe not.
 
and you only get 64 tabs and line feeds each
 
9:52 PM
Yeah.
Obviously interesting, but not the ideal challenge.
I should invent a language with @Doorknob this summer and some other people this summer that is top-notch at this.
 
I just had a brilliant idea, if we get to 1 character left I am prepared. — Pietu1998 7 hours ago
 
Something along the line where you can write without any given combination of letters. Sounds impossible though.
 
^ That'd probably do it. :P
 
OK, that is for the next one.
@Pietu1998 Unfortunately, I think the general rule for this sort of challenge is that you can't use languages made up after the challenge began...
Sorry for the ping, Pietu.
 
@Doorknob sweet
 
9:59 PM
@hosch250 (S)he won't get it anyway, since pings only work for people who have been in the chatroom in the past {something something units amount of time}. If the name autocompletes in chat, then the ping will work.
 
OK. That will break the answer form.
Did you see how long the Hello World program is? It would probably crash my 8-GB RAM computer!
 
@hosch250 You'd just submit the length. Have a look at the answer in Unary.
 
OK...
That language can never be blocked then.
Might as well start my calculation now.
Or maybe not.
On a different note, have any of you ever received more than 100% on an exam?
 
Uh oh. Is PPCG being borky for anyone else right now?
llama@llama:~$ ping codegolf.stackexchange.com                         17:08:46
PING codegolf.stackexchange.com (198.252.206.16) 56(84) bytes of data.
64 bytes from stackoverflow.com (198.252.206.16): icmp_seq=1 ttl=50 time=182 ms
64 bytes from stackoverflow.com (198.252.206.16): icmp_seq=3 ttl=50 time=122 ms
64 bytes from stackoverflow.com (198.252.206.16): icmp_seq=5 ttl=50 time=131 ms
64 bytes from stackoverflow.com (198.252.206.16): icmp_seq=6 ttl=50 time=126 ms
64 bytes from stackoverflow.com (198.252.206.16): icmp_seq=8 ttl=50 time=119 ms
I can't seem to access the home page
 
I have; I got one problem wrong (or was it two?), but because there were more questions in the test than were supposed to be the calculation came out funny (this is Moodle, the glorious, open-source Uni software).
Moodle has lots of bugs right now.
@Doorknob I can get to it.
When did you become a mod on Puzzles?
 
10:13 PM
@hosch250 Yep, it just went back up. That was short. :P
 
I wouldn't want to be a mod, you can't participate as much.
 
10
A: Who should our moderators be?

DoorknobI'll go ahead and nominate myself as a pro-tem moderator for Puzzling. This site: Other sites: Why do I want to be a moderator? What would I do as a moderator? this is an interesting topic to me and many others that I know, and I'd really like to help the site succeed I will...

@hosch250 Why would you say that?
@hosch250 Whoops, wrong post. Here's the announcement one
 
Because the mods are looked on differently.
They can make official statements.
 
@hosch250 I've received 100% on an exam which I didn't complete. The wonders of adjusting to fit a curve.
 
Cool.
Thanks, Martin Buttner!
See you all later!
 
10:46 PM
1
Q: Running PHP with `-r` instead of code tags

chilemagicBased on this manual, there is a -r option that allows you to run PHP code on the command line without specifying the code tags, <? ?>. I recommended it on the polyquine question, where this answer code be considered a PHP quine if we used -r or -R. I'm pretty sure -R is used like php -R scriptn...

 
11:10 PM
0
A: Sandbox for Proposed Challenges

Beta DecayTetration Golf code-golf math Tetration, represented by na (or a^^n) is used to represent large numbers. Basically, it's a to the power of itself n times. For example: a^^n 2^^2 = 2^2 = 4 2^^3 = 2^2^2 = 16 3^^3 = 3^3^3 = 7.6255975e+12 The shortest code to output the result of na wins. A and ...

 

« first day (1369 days earlier)      last day (3480 days later) »