9
\$\begingroup\$

This is one area where I'm having trouble envisioning how this site is supposed to work as an SE. I suppose wrt the golfing there's an objective criteria--least characters. But for puzzles there's really not. Any answer that correctly solves the problem is technically an answer and so picking from among them seems to be impossible, objectively, unless you just grab the first one. That seems to kill the whole point of having a puzzle archived though, which would be for people to try and solve it themselves.

Even when there IS an objective criteria, someone could always come along and break the record. What are we supposed to do, set a time limit? Adjust the answer when someone comes along a year later with something better?

The problem I see here is that, fundamentally, puzzles and challenges are not questions. The person posting them may even have done it already. Very rarely does it seem appropriate to pick an "answer" to something posted not as a question, but a challenge.

\$\endgroup\$
12
\$\begingroup\$

I, too, don't see acceptance as meaningful for most puzzles and challenges.

Let the users vote and leave it at that.

\$\endgroup\$
  • 3
    \$\begingroup\$ I also agree. Should we remove the answer accepting feature from this site entirely? \$\endgroup\$ – marcog Jan 28 '11 at 18:24
  • \$\begingroup\$ @marcog: If we envision every question being either a puzzle, a game, or a poll it makes sense. If we're expected more traditional questions then probably not. \$\endgroup\$ – dmckee Jan 28 '11 at 18:30
  • \$\begingroup\$ What would you do if someone wanted to award a bounty? \$\endgroup\$ – Nick Larsen Jan 28 '11 at 19:02
  • \$\begingroup\$ @NickLarsen: I don't know...fume? There's not much I can do. \$\endgroup\$ – dmckee Jan 28 '11 at 19:24
  • 1
    \$\begingroup\$ @Nick I get the feeling if we choose to drop answer acceptance we might want to drop bounties as well. It's another one that time will tell though. \$\endgroup\$ – marcog Jan 28 '11 at 19:29
  • \$\begingroup\$ Code challenges that are particularly well done should deserve the option for awarding a bounty, like the twitter image encoder challenge or the battleship challenge. \$\endgroup\$ – Nick Larsen Jan 28 '11 at 19:30
5
\$\begingroup\$

The way I've run code golfs on SO is that I update the accepted answer to match the shortest answer at the time. So if someone posts a shorter answer, I update the accepted answer.

I concur with dmckee with regard to more-subjective questions that have no clear criterion of "best".

\$\endgroup\$
  • \$\begingroup\$ This is certainly reasonable. What do you do in the event of a tie? Earliest? Highest voted? \$\endgroup\$ – dmckee Jan 28 '11 at 19:24
  • \$\begingroup\$ Do we always want golfscript winning? See comments on this answer. \$\endgroup\$ – marcog Jan 28 '11 at 19:30
  • \$\begingroup\$ @dmckee: Good question. In the case of "stack overflow code golf" I went with "fastest gun in the west" (and certainly Anarchy Golf uses the same policy), but that may not be appropriate for every code golf challenge. Comments welcome. \$\endgroup\$ – Chris Jester-Young Jan 28 '11 at 20:05
  • 1
    \$\begingroup\$ @marcog: Contrary to urban legend, GolfScript is not suitable for every golfing problem. For example, no one has written a GolfScript solution for Decision Tree code golf, and even if someone did, it's very, very unlikely to beat the Perl solution. (The easy-path solution requires floating point and symbols, neither of which GolfScript has.) \$\endgroup\$ – Chris Jester-Young Jan 28 '11 at 20:07
  • 1
    \$\begingroup\$ @Chris That's great to know! I've never really looked into it. So the solution is to have more complex code golfs. :) \$\endgroup\$ – marcog Jan 28 '11 at 20:33
  • \$\begingroup\$ I wouldn’t mind you doing this, but I definitely would not recommend this as general practice because people forget and a longer answer stays accepted. It is a huge problem that the accepted answer can only be updated by the challenge setter, so it must only be done if the challenge setter is very likely to stay around for the (potentially unbounded) lifetime of the challenge. \$\endgroup\$ – Timwi Apr 3 '11 at 18:05
1
\$\begingroup\$

The lasers golf happened before I started golfing on SO. The accepted perl answer has since been beaten in Python, Ruby and Golfscript but LiraNuna always accepted the shortest answer 1 week after the question was answered

\$\endgroup\$
  • 1
    \$\begingroup\$ ...and at some part he started including the "end date" in his questions. That is, he posted them on Thursday and accepted the best answer Monday morning or some such. That's fair, but still not terribly interesting to me. \$\endgroup\$ – dmckee Jan 29 '11 at 15:57
0
\$\begingroup\$

Personally, I would think:

  1. On strictly objective questions, (eg. plain old code golf) the shortest answer at any given time.
  2. On questions with any subjective element, a bonus to the answer with the largest number of votes at any given time.
\$\endgroup\$

You must log in to answer this question.