11
$\begingroup$

Now that we can migrate off topic questions, what sort of migration paths should be useful to have?

I think that adding at least Cross Validated, if not one or two of the other science sites, would be reasonable.

(Note that there is a two months limit on migration. We won't be able to migrate posts older than two months.)

$\endgroup$
  • 2
    $\begingroup$ Re your edit: all the more reason to get the paths set up quickly. $\endgroup$ – David Roberts Jun 28 '13 at 10:08
  • 1
    $\begingroup$ Isn't there a severe limit to the allowed number of pre-programmed migration targets? Or have I mis-remembered something... $\endgroup$ – Gerald Edgar Sep 3 '13 at 12:07
14
$\begingroup$

Disclaimer: Hi! I'm one of the community managers at Stack Exchange. I'm sure you've seen some of us and a few developers around over the last couple of days, sorting out last-minute issues and problems arising from the site upgrade to SE 2.0. Welcome! We're happy to have y'all here.


Far as migration paths go, we'd be happy to set up up to 4 defaults for you guys, but there are a few considerations we usually take into account.

First of all, migrations paths are only available for graduated sites. This means that sites like Academia, Computational Science, Computer Science, and Philosophy are currently not an option for user-driven migrations. This is done to minimize the chances of newer sites and/or sites that have yet to cement their long-term success from being potentially overwhelmed with effectively rejected question from other sites in the network. We ideally want every beta to form its own voice.

In a pinch, diamond moderators have the ability to migrate a question to any site in the network (so long as it's not in the private beta stage), but this is ideally only done for exceptionally good questions and not just anything that could conceivably be asked somewhere.

Secondly, Mathematics is an obvious choice and it's already enabled, but beyond that I would like to suggest letting MathOverflow run as is for at least a few weeks to see what sort of questions you keep getting and what sites they get migrated to. We collect statistics on how many questions are migrated, where they end up, and whether or not those migrations are rejected by the destination site.

(Aside: migrations are considered to be rejected by the destination site when migrated questions are closed as anything other than "duplicate" or deleted. This is another flood prevention mechanism.)

In a few weeks/months, we all will have a better handle of what migration paths make the most sense and are most successful. In the meantime, you can request a migration by flagging a question and explaining what you'd like to happen, and a moderator will make the call whether to move it over to another site or not. We'll be happy to set up user-driven migration paths when we see clear patterns for migrated questions both in terms of quantity and quality (i.e. low rejection percentage).

$\endgroup$
  • 7
    $\begingroup$ Anna, I feel like we had this discussion before (and we did, on the meta.MSE site). If the migration paths are not open, there is a good chance people will prefer to just close it as off topic rather than flag it for migration. By that rationale, then, no migration path will be open because hardly any migration will occur. I think that opening migration paths for a month, or six weeks, and then closing those which are unused -- that's the better option. In that case, even if they are closed, users are still informed of those options and learn to flag for the migration. $\endgroup$ – Asaf Karagila Jun 26 '13 at 6:17
  • 6
    $\begingroup$ Anna, I think you should search the database of mathoverflow comments for the string "tex.stackexchange.com". I think you'll find that there are a reasonable number of cases where people have suggested that things should be moved there. Given that, I don't think we need to wait longer before setting up a proper migration path. I think that the same is true for stats.stackexchange.com. $\endgroup$ – Neil Strickland Jun 26 '13 at 9:42
  • 5
    $\begingroup$ I found 66 comments containing the string "tex.stackexchange.com", 50 of them on closed questions. Just FYI, @Neil. $\endgroup$ – Shog9 Jun 28 '13 at 3:02
  • $\begingroup$ @Shog9 Thank you for the number. Does this search include deleted questions? (In more recent times most such things got deleted, while before hardly anything got deleted.) The problem might be bigger than this. But in any case there is a fair number of off-topic question that might fit on TeX. They might not contribute much to the tex site but OP will still get much better information, even if perhaps only via the link for the duplicate vote. $\endgroup$ – user9072 Jun 28 '13 at 10:10
  • 2
    $\begingroup$ All comments on all posts, deleted and visible @quid $\endgroup$ – Shog9 Jun 28 '13 at 16:00
  • $\begingroup$ "migrations are considered to be rejected by the destination site when migrated questions are closed as anything other than "duplicate" or deleted." That doesn't appear to be true. If you look at the rejection stats for security->crypto they can only be explained if duplicates are considered "rejected". (6/14=42% vs. 2/14=14%) $\endgroup$ – CodesInChaos Jun 29 '13 at 10:29
  • $\begingroup$ @CodesInChaos There is some weirdness in how we calculate the rejection stats. You're seeing a 40% rejection rate on Crypto, but Security is showing 15%. However, that bug aside, closing as a duplicate doesn't reject a migration. When a migration is rejected, the question is locked on the destination site and unlocked/closed as off-topic on the source site. The redirect link between the two is also severed. This doesn't happen with questions closed as dupes on the destination site. $\endgroup$ – Adam Lear Jun 29 '13 at 15:53
  • 1
    $\begingroup$ Hi @AnnaLear, could we turn on the migration path to stats.stackexchange.com now? We've already moved a few across which were very clearly identifiable as stats questions, and I'd be happy if this could be done by our users rather than requiring a mod. $\endgroup$ – Scott Morrison Jul 9 '13 at 16:36
  • 2
    $\begingroup$ @ScottMorrison Done. $\endgroup$ – Adam Lear Jul 9 '13 at 16:42
6
$\begingroup$

The obvious answers : academia.stackexchange.com and math.stackexchange.com.

$\endgroup$
  • $\begingroup$ Well, I didn't write MSE because it's already open. But Academia is a very good mention! $\endgroup$ – Asaf Karagila Jun 25 '13 at 21:00
  • 3
    $\begingroup$ Plus, computer science, theoretical computer science, physics, philosophy,... $\endgroup$ – Andrés E. Caicedo Jun 25 '13 at 21:27
  • 5
    $\begingroup$ scicomp.stackexchange.com, mathematica.stackexchange.com/, tex.stackexchange.com $\endgroup$ – Neil Strickland Jun 25 '13 at 21:34
  • 2
    $\begingroup$ @NeilStrickland post that as an answer :) $\endgroup$ – Manishearth Jun 25 '13 at 23:21
  • $\begingroup$ David Roberts suggested statistics. (Which I just noticed was already mentioned by Asaf.) $\endgroup$ – Andrés E. Caicedo Jun 28 '13 at 2:41
  • $\begingroup$ But remember the limit of 4 sites. $\endgroup$ – Gerald Edgar Nov 20 '14 at 14:43
6
$\begingroup$

In my opinion besides math.SE the most important other site would be Academia, however this being impossible for the time being as explained by Anna Lear, my next three top-choices (I limit myself to three as there seems to be some concern regarding having too many open) would be (somewhat in order of relevance):

  1. TeX

  2. Mathematica

  3. Cross Validated

I agree with other contributors, especially Asaf Karagila and Neil Strickland, that there seems little reason to wait; there is enough historical evidence to know that such migrations are relevant. I do not see much reason this issue dissapears now just since we changed into the network (though it might help).

In any case, I think at least the path to TeX should be open as soon as possible; I think most instances there are also uncontroversial. This also applies mainly to Mathmatica, I think. For statistics questions there is more of a potential risk for controversy if a question or other should be migrated. But at the very least for TeX I would consider it as unfortunate if we would not get the paths now that it is possible; that now one could migrate questions was in advance of the move to the network in my mind a main point in favor of joining the network.

$\endgroup$
  • $\begingroup$ I am not sure, but IIRC there cannot be more than 3 migration paths. It would be better to post these separately so it is easier to evaluate the amount of support each of them has. $\endgroup$ – Kaveh Jun 27 '13 at 23:30
  • $\begingroup$ @Kaveh: in Anna Lear's answer it is said they would be happy to set up up to 4 defaults. So 4 seems like the minimum that is possible (and the formulation does not even suggest this is some strict upper-limit), which is why I suggested three in addition to the existing one. By post separately, do you mean I should post an answer for each so that people can vote on them individually? I am not sure this is necessary, but of course if somebody wants to do a poll they can go ahead. $\endgroup$ – user9072 Jun 27 '13 at 23:40
  • $\begingroup$ hmmm, I think she meant 4 as the maximum possible. On other sites Stack Overflow takes one of those slots, I guess that is why I thought max is 3. $\endgroup$ – Kaveh Jun 28 '13 at 0:37
  • 3
    $\begingroup$ @quid Four is the upper limit. Four other sites + the site's own meta (that is a migration path always available on every site). The minimum is zero. $\endgroup$ – Adam Lear Jun 28 '13 at 0:37
  • $\begingroup$ @AnnaLear: thank you for this clarification. This is not very much 4, but well. If 4 is really the maximum, perhaps let us keep one spare for the time when Academia will be mature and in the mean time we could have at least tex.SE in addition to math.SE (and perhaps Mathematica, or also something else). But, I think, really TeX is the most unproblematic one. If you want to give us an easy one, that is still useful, to practise, I think TeX is really the way to go. $\endgroup$ – user9072 Jun 28 '13 at 0:43
  • 1
    $\begingroup$ I don't think more than 4 is even remotely necessary since mods can always migrate anywhere else for the very rare exceptions. $\endgroup$ – François G. Dorais Jun 28 '13 at 14:06
  • $\begingroup$ @FrançoisG.Dorais: of course it is not necessary, until recently we did just fine for years without any migration path after all. But I think this is the wrong question to ask if it is necessary. Whether it is desirable or what vould even be reason against it seems more pertinent to me. But let us not discuss the number, as you see I suggested 4 and then reduced to basically just one in addition (tex) with the option of adding Academia later. Do you see a reason for having migration open to tex.SE? By contrast, I could see reasons why for stats.SE there could be reasons... $\endgroup$ – user9072 Jun 28 '13 at 16:30
  • $\begingroup$ @FrançoisG.Dorais ...to wait since there might be things to be discussed (on MO) and perhaps with the other site. But for tex.SE? And, I think to look at sheer numbers and to conclude this is not so frequent moderators can handle this. Time is a critical factor here. The faster the better. You (the moderators) will always remain able to migrate but to have an additional means for doing it seems like an improvement. (This is really not to complain, but it happened mod-flags are dealt with with significant delay only.) $\endgroup$ – user9072 Jun 28 '13 at 16:37
  • 2
    $\begingroup$ That was less precise than I wanted. Is it necessary to have a migration path to tex.se that does not require moderator intervention? I don't think so since I can't imagine we would suddenly get more than one tex.se every few months. At that rate, mods can do it manually upon request. $\endgroup$ – François G. Dorais Jun 28 '13 at 16:40
  • $\begingroup$ @FrançoisG.Dorais Finally, as a general addition. I am not for all kinds of paths, for example IMO Andres Caicedo's suggestion is a bit exagerated (and in the end unhelpful but perhaps it was only meant in a playful way). Specifically, I did not suggest TCS for example, for reasons close to the ones you explained some while ago on the metas. (And I assume you had not seen my second comment, regarding why the number of times this happens is nto so relevant. Also I asked if there is a reason against it if there is none why not have it, it is not as it costs a lot to open that path [I assume].) $\endgroup$ – user9072 Jun 28 '13 at 16:41
  • $\begingroup$ @quid, one issue with migration is that when a question is migrated it is closed and cannot be reopened anymore by users. This causes issues since other users might disagree with some migrations performed by other users, i.e. migration can be abused to permanently get rid of a question a few people just don't like. In addition migration needs agreement from both sites, migrating unsuitable questions would be annoying for the target site users. And there are a number of other issues. I think TeX - LaTeX is probably fine but for other sites the issue is more delicate IMO. $\endgroup$ – Kaveh Jun 28 '13 at 18:06
  • 1
    $\begingroup$ To sum all this up in general: if we can habe a path to math.SE, which we do, I see really no reason not to have one to tex.SE and perhaps mathematica.SE which is what I suggested. General concerns about potential conflicts might be better adressed to for example Andres Caicedo, suggesting all kind of things, not me. $\endgroup$ – user9072 Jun 29 '13 at 12:53
  • 2
    $\begingroup$ @FedericoPoloni I was rather in favor of having path open so I am not that well placed to reply. But, some problems: a) people not "understanding" the sites to which they migrate, actually migrating question someplace where they are off-topic, too. Of course this could be avoided if people just did not do that but well. b) migrating being used in as a "weapon" in open/close off/on topic "fights". The way some people suggested target sites to me suggestsed this was prevalent. It was more like "let us get rid of some 'nonsense' from MO" than "let's try to find a good place for some question" $\endgroup$ – user9072 Sep 3 '13 at 13:57
  • 2
    $\begingroup$ c) if there are too many possibilities it'd get tricky to find a consensus about the right site. Do we really want to decide here if a question is better for cs or tcs? This could be tricky. Alos, since that disucssion it came up that some people do not like this migration idea at all (and some of them vote to close a lot). So community-driven migration seems still more tricky to achieve if some do not participate in it (to be clear, I do not want to criticize those, but just point out the consequence). In any case, I think just path for the most frequent and rest via flags to mods is good. $\endgroup$ – user9072 Sep 3 '13 at 14:01
  • 1
    $\begingroup$ That being said, I agree scicomp is a very natural idea; I just did not discuss it as I knew it was impossible it being in beta. $\endgroup$ – user9072 Sep 3 '13 at 14:03
3
$\begingroup$

(This was a new question I asked, but François pointed out this should go here - no edits to the below)

Shouldn't we have migration to stats.SE at the very least? I've seen mentioned that physics.SE is an option, but this seems less urgent, and I've already seen a couple of questions that could/should have been there. For example https://mathoverflow.net/questions/135079/frechet-mean-of-multiple-curves - which is being voted to close with a custom reason that it should go to stats.SE.

$\endgroup$
  • $\begingroup$ (David, Cross Validated is the statistics site, it is mentioned by Asaf.) Also, this. $\endgroup$ – Andrés E. Caicedo Jun 28 '13 at 3:24
  • $\begingroup$ Andres - yes, but I thought it good to record question(s) which need said migration $\endgroup$ – David Roberts Jun 28 '13 at 4:10
  • 1
    $\begingroup$ +1 for giving an example. It strikes me as unfortunate that we still need to close questions and have people cross-post them manually while now it could be so much easier for everybody. $\endgroup$ – user9072 Jun 28 '13 at 10:05
3
$\begingroup$

I respectfully suggest http://scicomp.stackexchange.com as a new migration target. Sample questions for migration: 1 2 3 4 5. I have no access to deleted questions, so there could be more.

$\endgroup$
  • 2
    $\begingroup$ This site is currently beta, so we can't open a migration path to them. However, the moderators can still migrate questions there if they are not too old. Please flag the posts and indicate why & where they should be migrated in the comment box. $\endgroup$ – François G. Dorais Sep 2 '13 at 15:47
  • $\begingroup$ What kind of comments should I write on the "why" part? For most of the questions, I find it hard to make a different argument than "this should go on scicomp because it's about computational science rather than research math". $\endgroup$ – Federico Poloni Sep 2 '13 at 17:20
  • $\begingroup$ The question also needs to be suitable for the target site. So, in addition to being on-topic, it should also be a good question that won't be immediately closed for other reasons. Ideally, any additional justification that would help the moderators evaluate this should be included. $\endgroup$ – François G. Dorais Sep 2 '13 at 17:30
  • $\begingroup$ I've looked through these questions. Some are too old to migrate, some received some okay answers here already so I think it's best to leave as is. $\endgroup$ – Scott Morrison Sep 3 '13 at 0:25
  • $\begingroup$ I agree, however, that if flagged for migration immediately some of these would have been better sent over to scicomp. $\endgroup$ – Scott Morrison Sep 3 '13 at 0:26
1
$\begingroup$

Now that the Mathematics Educators SE is operational, a migration path there could be useful. This question could be migrated there, for example.

$\endgroup$
  • 2
    $\begingroup$ That site is still in beta. As far as I recall, it is not possible to get migration paths to beta sites. $\endgroup$ – Tobias Kildetoft Nov 19 '14 at 8:17
  • $\begingroup$ @TobiasKildetoft, I didn't know of that restriction, thanks. If that is indeed the case, we need to wait until it gets out of beta. $\endgroup$ – Joonas Ilmavirta Nov 19 '14 at 10:11
  • $\begingroup$ As @TobiasKildetoft says for now there cannot be an open path. But moderators can still migrate. You could flag "other" and suggest migration. $\endgroup$ – user9072 Nov 19 '14 at 18:48

You must log in to answer this question.

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