15
\$\begingroup\$

The reason for this post is due to the binary decimal representation challenge, where @primo is, as of writing, awarding a bounty for the fastest solution to a code-golf question.

In response to this, I'd like to raise a few questions:

  • Is it acceptable to start a bounty for a question with an alternative winning criteria? (e.g. "least whitespace characters used", "most popular" or "least memory used" to a code-golf question)

  • If the answer to the first question is yes:

    • Should the new scoring criteria be somewhat related to the original scoring criteria for it to be acceptable? (e.g. "least whitespace characters used" is tangentially related to code-golf, but "fastest code" is basically a polar opposite)

    • Should an answer aiming for the bounty also try to conform to the original scoring criteria even if it doesn't make much sense to do so? (e.g. golfing code intended for a "fastest code" bounty)

  • If the answer to the first question is no:

    • Is it acceptable to form a related question with the new scoring criteria, possibly with the permission of the original poster? (assuming this does not cause the new question to be a dupe)

    • Are there any alternatives to the previous question? (e.g. if we're desperate to see an efficient implementation for a code-golf)

    • Is there anything we should do if we see an alternative winning criterion bounty?

\$\endgroup\$
5
\$\begingroup\$

Yes and No

There are two kinds of bounties we can disallow:

  1. Bounties that reward answers that don't fit the submission requirements: "Draw a cat" on a Fibonacci challenge.

  2. Bounties that have a tendency to push answers away from the scoring criterion

For example, on a posting a bounty for "Fastest code" would not be allowed. Posting a bounty for "Code that is fast and golfed" is allowed.

Note that objectiveness is not required for bounties. However, "Fast code", "Well-formatted code", "Code with good variable names" would all be disallowed on , as they nearly always produce un-competitive answers.

However, if the bounty has little effect on the competitiveness of answers, or it takes into account the current scoring criterion, then it is fine.

\$\endgroup\$
  • 1
    \$\begingroup\$ I would argue that a bounty for "fast and golfed" still conflicts with the winning criterion of a golf challenge, as it rewards an answer that is partly but not optimally golfed. It encourages paying lip service to the winning criterion rather than optimising for it. \$\endgroup\$ – trichoplax Oct 25 '16 at 11:57
  • 1
    \$\begingroup\$ I'm not against combining winning criteria in a challenge, but that should be defined in the challenge wording. If a bounty encourages not optimising for the winning criteria defined by the challenge then I see it as paying to reduce the quality of the submissions. \$\endgroup\$ – trichoplax Oct 25 '16 at 12:01
  • \$\begingroup\$ @trichoplax to an extent, there are many bounties that have rewarded answers that aren't perfectly optimized. Consider a challenge that requires answers to do X, but the bounty would reward answers that can also handle the edge case Y. Answers would be golfed, but not necessarily as golfed as possible (to do only X) \$\endgroup\$ – Nathan Merrill Oct 25 '16 at 12:43
  • \$\begingroup\$ I agree we have had such bounties in the past. I think they should be ruled out in future. \$\endgroup\$ – trichoplax Oct 25 '16 at 17:33
  • \$\begingroup\$ @trichoplax then, effectively, the only bounties I can offer is "The best solution" or "The best solution in language X"? \$\endgroup\$ – Nathan Merrill Oct 25 '16 at 17:40
  • \$\begingroup\$ I'm trying to think of other bounty reasons that wouldn't conflict. It makes me realise that a golf often has several submissions in the same language using different algorithms, and I like seeing an interesting approach even if it doesn't turn out to be the shortest. This is hard to judge... \$\endgroup\$ – trichoplax Oct 25 '16 at 18:21
-1
\$\begingroup\$

No.

Bounties should reward exemplary solutions, if relating to an answer at all. i.e. shortest solution to a code-golf, as it is the best solution at the moment, fastest to a fastest-code, etc.

If you look at the bounty options, the only 4 really relating to answers are "Reward existing", "Canonical answers required", "More detail", and "Answers are out-dated."

None of these say anything about helping with a different problem. The closest might be "more detail", but that doesn't really fit.

\$\endgroup\$
-2
\$\begingroup\$

Yes

You should be free to create any bounty you like, as long as the OP is okay with the bounty.

\$\endgroup\$
-6
\$\begingroup\$

Yes

But the bounty scoring must have been used on this site (so fastest code is okay, but least whitespace is not) and new answers should aim to maximize their score according to the challenge, not the bounty.

\$\endgroup\$
  • 3
    \$\begingroup\$ I don't see any reason to put such restriction. Giving a bounty can have many reasons, and awarding a creative submission should be allowed. \$\endgroup\$ – Mhmd Oct 25 '15 at 15:04

You must log in to answer this question.

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