5
\$\begingroup\$

I'm as guilty (perhaps even more so) as the next person of this, and perhaps that's why I feel this question needs to be asked.

Often times, I'll come across a question which already has some answers - usually - which are really well-suited to the challenge requirements and winning criteria. For most cases, I would be hopelessly out-classed in the competition as it currently stands because my coding experience is fairly limited to begin with and most definitely does not include any languages designed with golfing in mind.

However, I'm still interested in the challenge and would like to effectively pit my wits against those of other people who write in my language. So, despite the fact that I cannot hope to win the challenge myself, I'll post my answer anyway.

Generally speaking, this seems counter to the regular StackExchange practice where new answers should really only be posted if they're contributing further towards the goal of the question. If someone's got a 20-character solution (albeit in another language), and mine is upwards of 100 characters with little to no room for reduction, my answer does nothing really except to add to the pile-up of losers.

This also, after a period of time, results in some challenges - even though they may have been won very early on with extremely short answers - garnering enough late-coming losers that it kicks over to Community Wiki and forces everyone who participated up to that point to lose their rep. The CW trigger is arguably a separate issue entirely, which is being addressed in other Meta posts. However, this particular behavior is a noteworthy contributor to that problem.

What should we be doing in cases like this? Should we only post an answer if it can actually beat the existing solutions? Or is posting a known-losing solution generally intended to be acceptable?

There are two cases where I think we can all agree that posting a known-losing solution is definitely not acceptable:

  • Your solution is not significantly different from all existing solutions in the same language.
    • For this case, comment on or propose an edit to the existing solution that is the closest match to yours.
  • Your solution would not win against the best currently existing solution in the same language.
    • For this case, just don't post it at all.

Aside from those two, should we still be posting losers?

\$\endgroup\$
7
\$\begingroup\$

It's acceptable to post "non-winning" answers in other languages, because otherwise, nobody would be allowed to post anything after the first GolfScript or J or APL submission, and that would be sad. :'(

More seriously, some people like to do a per-language leaderboards, so enabling people to post other-language solutions, even if longer than a GolfScript solution, would help with that.

\$\endgroup\$
  • 5
    \$\begingroup\$ The more I get involved here, and especially the more I read Meta, the more I keep thinking this site needs loads of features that other StackExchange site's don't. In fact, I sometimes worry that could be a big factor keeping it in beta. Prime example here is that we could use built-in support (and better reward systems - e.g. multi-acceptance) for per-language rankings within a thread. \$\endgroup\$ – Iszi Dec 12 '13 at 21:45
4
\$\begingroup\$

In addition to Chris' comments on the value of per-language competition, there is value in solutions that take a completely different route to solving a problem even if that solution is not shortest.

Note that for there to be a lot of value in this kind of answer, the post should include a ungolfed version of the code and some explanation of the method. But then, I think, that every post that is not trivial should include those kinds of luxury features.

\$\endgroup\$
-1
\$\begingroup\$

I've felt the same dilemma, and while I don't really have a solution, there is a third option: CodeReview.SE.

When I stumbled upon the Brainfuck challenge, I excitedly posted my program. But some time later ... looking over it again ... I knew it was the wrong place. I simply had no intention to golf it at all.

original (deleted)
re-post

\$\endgroup\$
  • \$\begingroup\$ No, I don't think Code Review SE is for asking people to post different-language implementations of your challenge. It's for you to post working code and get review comments on it. \$\endgroup\$ – Chris Jester-Young Dec 19 '13 at 20:42

You must log in to answer this question.

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