I noticed a couple of answers to the tweetable art question have been converted to community wiki. These are questions which, although interesting, didn't keep to the question requirements (the code was longer than 140 characters and additional functions were added outside the specified red, green and blue functions - which the question forbids).

I can understand why people would be frustrated at seeing answers that don't match the question, especially if they spent a long time golfing down their own submission to 140 characters... It doesn't seem a fair comparison.

However, I don't understand the approach of making these community wiki. That means the rep doesn't go to the answer poster, but the valid answers are still subject to a comparison which is not like for like, which may affect how they are received. At first glance it still may not be obvious that the invalid answers don't fit the question.

What approaches make sense in such circumstances?

Is there a benefit to making them community wiki that I'm missing?

  • 3
    \$\begingroup\$ Good question. The main point was not to rake in rep for an answer that doesn't meet the spec. But probably deletion would be the more reasonable alternative. Although I'd hate to see the Mandelbulb go. \$\endgroup\$ Aug 6, 2014 at 13:11
  • \$\begingroup\$ Is there a way to leave a stub rather than completely delete? So there would be a link saying "this was not a valid answer but follow this link for the interesting content and image"? \$\endgroup\$ Aug 6, 2014 at 13:18
  • 4
    \$\begingroup\$ I don't know... technically all they are is "Not An Answer". I can create wonderful images if I throw out the restrictions. And I think linking to something like that as supplementary material is fine if you have related valid answer (like I did with the random paintings). But a post just containing that... at best a link in a comment on the question. \$\endgroup\$ Aug 6, 2014 at 13:21
  • \$\begingroup\$ Perhaps they should be deleted as answers and reposted as questions to see if anyone can golf them down to the required limit... \$\endgroup\$ Aug 6, 2014 at 13:24
  • \$\begingroup\$ I just downloaded the mandelbulb image - it looks even better in full resolution. I still think it needs to go though. \$\endgroup\$ Aug 6, 2014 at 13:27
  • 2
    \$\begingroup\$ I was under the impression that the idea was to avoid the "oh, I was working on this for a while, but it just barely doesn't fit the spec, aww, I did that for nothing, and it looks really nice"-type thing, but I'm all for whatever everyone else thinks about this \$\endgroup\$
    – Doorknob
    Aug 6, 2014 at 13:28
  • 1
    \$\begingroup\$ It's like entering a three legged race on a motorbike. \$\endgroup\$ Aug 6, 2014 at 13:28
  • \$\begingroup\$ My first reaction was to try and find a way to golf it down so I could suggest it in the comments, but it's beyond me. \$\endgroup\$ Aug 6, 2014 at 13:29
  • \$\begingroup\$ I'm going to delete mine now. Might undelete if I can be bothered to fit it into the spec. \$\endgroup\$ Aug 6, 2014 at 13:58
  • \$\begingroup\$ Might well be possible now with the shortened function names and forward declaration... \$\endgroup\$ Aug 6, 2014 at 14:19
  • \$\begingroup\$ There's a precedent of answers saying "This doesn't exactly fit the spec but I'm sharing it because it's cool" being posted and getting upvotes in everything from popularity contests to code golf. A decision on this would have to be a broader policy. Personally, I don't see the harm as long as upvoters know what they're upvoting. \$\endgroup\$
    – xnor
    Aug 7, 2014 at 13:23
  • \$\begingroup\$ The only time "comparing" answers is a concern is in a popularity contest. In a speed / golf / etc contest it doesn't matter how well an entry does that breaks the rules, it won't get accepted or bountied. Perhaps something we might want to aim for is some standard boilerplate at the top of an answer to indicate that it's not acceptable? If it's CW then it's even easier to add it, rather than relying on the poster to add it. \$\endgroup\$
    – Sparr
    Aug 10, 2014 at 14:21
  • \$\begingroup\$ It's also worth considering that a borderline answer might prompt the question to be changed because the asker agrees with the answerer that the answer should be within bounds. Case in point, the forward declarations of the functions in the tweetable art question were added after an answer needed it, and everyone seems to agree it's a reasonable change. If we remove borderline answers, they might not prompt such changes. \$\endgroup\$
    – FireFly
    Aug 12, 2014 at 19:20
  • \$\begingroup\$ @FireFly I agree for borderline cases - I think this site is all about borderlines... What I want to see deleted are half finished answers or answers making no attempt to fit the question. The answer linked to just defines new functions which means anyone could write any arbitrary program. I wouldn't want borderline cases to be swept up with these. \$\endgroup\$ Aug 12, 2014 at 19:25

2 Answers 2


Having given this some thought, I think the best approach is to both downvote and flag such answers for moderator attention. Downvoting alone has no noticeable effect when the people upvoting the image outnumber the people downvoting the code.

In this case I did flag and the moderator response was to convert to community wiki, as that was what was suggested by others at the time. Having taken time to think about this I believe this is an inconsistent approach and is unfair on the other answers.

There was an answer that was deleted

It was a picture of a cat with code that clearly did not produce it. Another answer with an image more impressive than most but only because it was produced with code not even close to following the rules has instead been converted to community wiki and massively upvoted. I think this inconsistency sets a precedent that needs to be dealt with now. It doesn't matter that the rep does not go to the poster of the answer. The answers should all contain images relevant to the question so that they can be compared on an equal footing. Otherwise this site is just a image repository.

In any question (particularly but not only popularity contests) there is the possibility of answers being eye catching but unrelated to the question, and getting net upvotes regardless of how many people downvote. The reason we need moderators is because sometimes downvoting doesn't work. The picture of a cat had net upvotes when it was deleted by a moderator. I think the same should apply equally to all questions that do not attempt to stay within the rules.

Answers that beg to get around the rules because "I don't want to waste all that effort" should be deleted to set a clear precedent. They should be undeleted only if they can later be modified to work within the rules. For this reason I believe the answer linked to should be deleted.

If you disagree, please post an answer stating and explaining your objection so there can be a consensus to give clear guidance for future cases.

  • \$\begingroup\$ I'm also concerned that there is no incentive for someone to come back to a half finished answer if they can just post it as is and see it voted to the top. Deleting such half answers until they are ready to be posted improves the content on the site. \$\endgroup\$ Aug 12, 2014 at 19:08

If we are considering a significant amount of moderation power being applied to enforce this rule, I think it is important to ask who this behavior is harming.

PCCG is not a normal SE site. The average visitor is not coming here to find the "answer" to a "question". No one is going to suffer from following a "wrong" "answer" by finding a highly upvoted answer here that doesn't strictly follow the rules of the question.

Upvotes are [mostly] not finite. One wrong answer getting more upvotes does not mean that less-interesting more-valid answers will get fewer. I can upvote the pretty picture AND upvote the best solution to the challenge within the rules (and the one in my pet language, and the one with the most vowels, and...). The best valid answer will still get accepted / bountied. Good valid answers will still get upvotes.

At most, I would like to see us adopt some standard boilerplate (a Template, in wiki terms) to put at the top of invalid answers. When they are community wiki, this will be trivial to implement. When they aren't, edits will have to be approved.

  • 3
    \$\begingroup\$ I don't think we're considering applying a significant amount of moderation power now to enforce the rule, but if we communicate the message that non-answers are acceptable then they will increase until it is necessary for the moderators to do a lot of work to get things back under control. We've seen it already once this year with code trolling, which ended up requiring an awful lot of effort simply because no-one was around to put a lid on it when it started. \$\endgroup\$ Aug 10, 2014 at 14:57

Not the answer you're looking for? Browse other questions tagged .