26
$\begingroup$

We are regularly migrating non-research-level questions from here to CS.SE. I think it would be good to open the migration path to CS.SE so high reputation users can choose to migrate to CS.SE when closing a question (without moderator intervention).

What do you think? Should we go ahead and ask SE to open the migration path to CS.SE?

What about the migration path to MSE?


Update 2015-08-23:

CS.SE held its moderator elections recently and it is now possible to set a migration path to CS.

$\endgroup$
  • 2
    $\begingroup$ Yes, yes, yes ! $\endgroup$ – reinierpost Apr 11 '13 at 8:33
5
$\begingroup$

Now that both sites have graduated, let's revisit this request. I proposed some criteria last summer (which we've been using internally for a while). My evaluation:

  1. Sanity check: Yes, there are plenty of reasons a question might be asked here that really belongs on the more general CS site.

  2. Gather statistics: In the past 90 days, there have been 6 migrations to and from CS Theory:

    cs.stackexchange.com          => cstheory.stackexchange.com × 4 (25% rejected)
    datascience.stackexchange.com => cstheory.stackexchange.com × 1
    cstheory.stackexchange.com    => cs.stackexchange.com × 1
    

    This is underwhelming. It seems that in practice, people are pretty good about posting to CS when they have a non-research-level question. Just one question has been migrated away. It does not seem as if the moderators are being overwhelmed with migration flags. It's also instructive that there are few migrations the other direction. (A 25% rejection rate would be a red flag if it were not just one question.)

So I don't see a lot of evidence that this path is needed. Looking at the close statistics it seems that the custom close reason that directs people to post to CS on their own is the prefered solution. 115 questions have been closed that way over the same time period. Now if the proportion does shift toward migration and if those migrations are successful as a rule, we might have a reason to add this default path.

$\endgroup$
  • $\begingroup$ Jon, we typically close off-topic questions without migration. Migration needs a mod action and mods very seldom do so. I think the real criteria is how many off-topic questions would be properly migrated. Also note that 25% is just one question, i.e. the sample size is way too small for a conclusion about the probability of improper migrations is high. $\endgroup$ – Kaveh Dec 23 '15 at 1:56
  • $\begingroup$ @Kaveh: Makes sense. If you feel like experimenting with migrating questions rather than straight up closing them, we can revisit in a month or two. $\endgroup$ – Jon Ericson Dec 23 '15 at 19:32
  • $\begingroup$ So the migration path is going to be activate temporarily for a few months to collect better statistics? Sounds good to me. $\endgroup$ – Kaveh Dec 23 '15 at 21:27
  • 2
    $\begingroup$ @Kaveh: That's not what I meant. Instead, I suggest encouraging people to flag for migration and/or moderators manually migrate questions that have been closed with this custom off topic reason. It's likely there will be plenty of good migrations to turn the default path on, but I'd like to be sure. (In case you hadn't read Shog's answer, migrations are a feature we hope to de-emphasize. They are often a poor user experience for everyone involved.) $\endgroup$ – Jon Ericson Dec 23 '15 at 21:35
  • $\begingroup$ I doubt that would work for the same reasons that people do not normally flag for migration and moderators seldom migrate questions by themselves. If you want a better estimate of how often the migration path could be used if we had it you can go over the questions closed as off-topic and see how many of them would be on-topic on cs. My impression is that we get such questions quite regularly. $\endgroup$ – Kaveh Dec 23 '15 at 22:37
  • 1
    $\begingroup$ Rather than training people to flag if they're not currently doing so, I'd recommend just periodically reviewing closed questions and migrating those that warrant it, @Kaveh. This gives you a chance to test two things: how many off-topic questions are actually worth migrating, and how many get rejected in practice. No sense training folks to pan for gold if there isn't any. $\endgroup$ – Shog9 Dec 23 '15 at 22:47
20
$\begingroup$

Speaking as a user on cstheory.SE: yes, please! Finally there is a way to treat good but low-level questions properly, we should use it. Several of the high-rep users here are also active on cs.SE so we have both implicit (we are all (?) computer scientists) and explicit expertise regarding the scope of cs.SE.

Speaking as a mod on cs.SE: Sure, keep them coming -- as long as the questions are good! I trust the cstheory community to know and respect our scope. So far, cs.SE has clearly profited from the migrations. I assume the volume will drop once cs.SE becomes more visible and more known in the SE community.

cstheory -> cs is probably the most natural migration path there is in the current SE network. In the future, cs -> cstheory will be used, too; we have had several theory questions which have remained unanswered. More on that later.

Update: As a mod on cs.SE, I'm still in favor of this migration path. We have enough community moderation in place by now to effectively and efficiently deal with cases of accidental crap migration.

$\endgroup$
7
$\begingroup$

Since we have high-rep CS.SE folks here, I don't see a problem with doing this. But I'd like Raphael and Gilles to weigh in as well.

$\endgroup$
4
$\begingroup$

As a rule, we don't open migration paths to or from sites in beta. This makes a lot of sense, so I'm glad to see the support for it - once CS.SE graduates, we'll revisit this.

Until then, folks should flag good but off-topic CS posts for moderator attention, as moderators can migrate anywhere.

$\endgroup$
  • 3
    $\begingroup$ Just the frequency of migrations is sufficient reason to take the site out of beta, right? $\endgroup$ – reinierpost Apr 11 '13 at 8:34
  • $\begingroup$ Not by itself, no - the site needs to be able to sustain itself without migrations as well. But there are certainly enough migrations happening to warrant a path once it does graduate. $\endgroup$ – Shog9 Apr 11 '13 at 14:32
  • 2
    $\begingroup$ @Shog9, can we now have the migration path set? :) $\endgroup$ – Kaveh Aug 24 '15 at 2:52
  • $\begingroup$ It's been 6–8 weeks since CS graduated. 3 years since both communities have made the request. How much longer? $\endgroup$ – Gilles 'SO- stop being evil' Dec 23 '15 at 0:29
  • $\begingroup$ @gilles, see Jon's answer. He is saying that there isn't enough evidence in terms of migrations to support the request for a migration path from cstheory to cs. $\endgroup$ – Kaveh Dec 23 '15 at 22:33

You must log in to answer this question.

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