9
$\begingroup$

Inspired by this question, and partially addressed in this meta post.

First off, the question linked isn't closed, and may never be, so this is more of a general question cum proposal.

Should there be a feature allowing us (probably meaning the moderators) to move good answers from questions marked duplicate to be answers of the original question? My argument for is that concentration of good answers improves the experience of outsiders who might stumble onto the site (i.e. through google).

For discussion, this strikes at something that has been a pet peeve of mine for some time. If an answer cannot be moved to the original question because it doesn't make sense, doesn't that indicate that the supposed duplicate is not one at all? AndreiROM's answer to the question I'm talking about would make perfect sense as an answer to the question proposed as a duplicate, so that implies that the assignment as a duplicate is valid.

$\endgroup$
6
$\begingroup$

If every answer on the duplicate is also a valid answer on the original, then it makes sense to move them. Sometimes the duplicate is more specific than the original, and answers on that duplicate don't cover all the cases in the original; in that case I don't think it makes sense to move. It also night not make sense if both questions have collected approximately the same set of (duplicate) answers, which sometimes happens.

So before taking this action, somebody needs to review both questions and all answers to determine if they are good candidates for consolidation. It would be best if such review came from the community, to both share the work and tap the right experts.

Moderators can merge questions, which has the effect you're looking for. A merger leaves one question with all the answers and the other question as a merge stub with a link to the other question. Because the questions are true duplicates, the "surviving" master question can be either, so we'd prefer to take the one that is better, whatever that means in a particular context. A question can also have only one accepted answer, so if both questions have accepted answers at the time of the merge, one of them is going to lose that status. So, again, community input is helpful here.

I suggest using meta to assess individual cases. If you see a true duplicate with answers that would benefit from being all in one place, bring it up and see what the community thinks about merging them. Try to get consensus on the decisions that need to be made. Then ask a moderator to pull the trigger.

Thus far we have had only two mergers on this site (1, 2), and they were exact duplicates asked by the same user. But other sites have had mergers that were asked by different people with different wording, so this is something that is done sometimes. Mi Yodeya, for example, has 72 mergers out of 21k questions. Meta.SE has 81 mergers out of 81k questions.

$\endgroup$
4
  • $\begingroup$ I did not know mods could merge questions. Has that ever been done before (on this site)? $\endgroup$
    – kingledion
    Mar 10 '17 at 22:38
  • $\begingroup$ @kingledion hmm, good question -- I don't see a way to find them if so. (And I've checked the mod tools on another site where I know we have merged questions, so it's not a case of "they'll show up if you have them but you don't".) $\endgroup$ Mar 10 '17 at 22:41
  • 4
    $\begingroup$ Found it -- we've had two mergers: worldbuilding.stackexchange.com/questions/59726/… and worldbuilding.stackexchange.com/questions/3116/… are the merge stubs. $\endgroup$ Mar 10 '17 at 22:56
  • $\begingroup$ Nice finds. It might be worth noting that both of those were merged because 1) Someone asked the same question twice on the site, or 2) Someone cross-posted, the version on the other site was sent here, and they were merged accordingly. $\endgroup$
    – HDE 226868 Mod
    Mar 10 '17 at 23:00
0
$\begingroup$

Makes sense
and for one simple reason: As it stands, if someone comes in from Outside and finds the "closed as duplicate" query, there will be a link to the question this one is a duplicate of, so our intrepid researcher will have access to both questions and both sets of answers. However, if she should chance upon the "original" / non-marked-as-duplicate query, there will be no such link to any of the duplicate queries. She'll only have access to that one question and its answers.

Merging two queries that are determined to be actual duplicates has the advantage of putting all the eggs in one yummy basket of geopoetical goodness. There's no need to link back to the merged question, because it won't have any answers to miss out on.

So this meta discussion begs two questions: why don't we merge queries more frequently? and also: can more than one duplicate query be folded into one? (For the latter, I suppose by serial merging.)

As for the former, this seems like a reasonable practice to engage in. As a community, we already frown on people asking duplicate questions, and for good reason. It seems to me that marking as duplicate is but a half measure: why leave the question in duplicate limbo when it can be profitably combined with its predecessor?

The only down side: we do sometimes argue about whether a query is really a duplicate or not. Perhaps merging could be done after a discussion in meta?

$\endgroup$

You must log in to answer this question.

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