4
$\begingroup$

I realize that there has been a lot of discussion about cross-posting an migrating between MO and MSE. It seems to me that it is common for someone to be unsure on whether a question is an MSE question or an MO question. An appropriate solution in this case is to post the question on MSE wait a good period of time. Then, if there is no activity on MSE, this inactivity can be part of a justification of a cross post on MO. I don't know if there is a consensus on what a good period of time might be, but it seems to be at least a few days. With this in mind I wanted to ask the following two questions:

1) Of questions that get answered on MSE, does anyone have data on how long it takes for 50%,60%,70%, etc. of questions to get answered? Or alternatively, how what percentage of questions get answered after 48 hours? 72 hours? 96 hours?

I confess that the above question is a bit leading and would require data from a site that is not MO. However, it's my opinion that a person should wait a week before cross posting. Admittedly this opinion is based on no evidence, and so having this data might help inform an opinion on that appropriate time to wait before cross-posting and better frame such a discussion. With this in mind I wanted to propose the following feature:

2) Should MO add a feature that flags questions having phrases similar to "cross-posted", "MSE", or "math.stackexchange" ask the person posting to provide a link to the question? I suppose the feature could automatically check the time the question was posted on MSE and provide data on how long it usually takes to get an answer on MSE. In the absence of such data, perhaps this feature could just set a waiting period of at least $n$ days.

I realize that this agreement on the value of $n$ might be a contentious. However, I just saw a cross post where the waiting period was 7 hours. I feel like this kind of behavior should be discouraged.

$\endgroup$
8
  • 1
    $\begingroup$ Probably makes more sense to ask your first question at meta.mse (or to search meta.mse for an answer, as it may have come up already). My impression is that if a question on m.se hasn't been answered in 24 hours then it has been buried so deep under the avalanche that it's fairly unlikely to be answered, ever. The 7-hour waiting period already is discouraged, in the sense that usually someone comes along to rap the offender's knuckles. The real problem (I think) is the folks who cross-post without telling anyone at either site. $\endgroup$
    – Gerry Myerson
    Oct 9, 2014 at 4:38
  • 1
    $\begingroup$ I'm not sure if the answer to your first question would be of much help. Most questions at MSE are simple homework. The hard questions for which MO might be an option are a minority, and what we need here is data on those questions only. I suppose the typical answering delay is much longer for the hardest MSE questions. $\endgroup$
    – Joonas Ilmavirta
    Oct 9, 2014 at 7:27
  • 4
    $\begingroup$ Honestly, I feel that all cross-posts should be discouraged. (SE itself strongly discourages any cross-posting.) Instead, users can flag their math.SE questions for migration to MO. (Right now there is no migration path from math.SE to MO, and so such migrations can only be made by math.SE moderators. With a joint moderator between the two sites this can be smooth, and we do look closely at each such flag, of which we receive about one or two each week.) $\endgroup$
    – user642796
    Oct 9, 2014 at 8:50
  • $\begingroup$ Here are the response time statistics for math.se questions given that the answer took more than 6 hours. data.stackexchange.com/math/query/234500/average-answer-time I'm not sure what statistical model makes sense but the $1.65\sigma$ rule-of-thumb suggests a 4 month wait to ensure 5% error... Anyway, the query can be modified easily to get more relevant data. $\endgroup$
    – François G. Dorais
    Oct 14, 2014 at 16:31
  • $\begingroup$ @ArthurFischer What is the claim that SE strongly discourages any form of cross posting based on? If SE would actually have a "each question only once in the network" policy it seems they ought to have cross-site dupe closure as option, which they do not. $\endgroup$
    – user9072
    Oct 16, 2014 at 18:23
  • $\begingroup$ @quid: Consider this MSE post that has been raised to (faq) status. It seems to have been accepted that some forms of cross-posting aren't necessarily bad (for example, and old unanswered question that can no longer be migrated), but these seem to be exceptional cases. (I believe that the data for different SE sites is stored in different databases, which would make searching for cross-posted questions more resource intensive. This would be one reason SE doesn't actively search for cross-site duplicates as they do in-site duplicates.) $\endgroup$
    – user642796
    Oct 17, 2014 at 2:54
  • $\begingroup$ @ArthurFischer I mainly reacted on the "any form"; had you said "most forms" or something like it right away I likely would not have commented. But, also, the following is quite close to my opinion meta.stackexchange.com/a/16708 I do not think it is only a technical issue, no tools for this are in place. For a well received on topic question I favor crosspost with links over migration; migration is deletion down the road (on source), we do not even delete on site dupes but leave them as signposts, why not have the crosssite "dupes" in the same spirit. $\endgroup$
    – user9072
    Oct 17, 2014 at 10:18
  • 1
    $\begingroup$ @quid: On re-reading, my original language was too strong by half. (I've grown very accustomed to being able to edit comments whenever I want/need to!) On math.SE we get a fair number of questions which are crossposted within hours (if not minutes) to/from physics.SE, cs.SE, SO, and other sites without any mention that this has occurred. It frustrates me when this is eventually brought to my attention, and my comment incorporated too much of this frustration. $\endgroup$
    – user642796
    Oct 17, 2014 at 10:22

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .