11
$\begingroup$

A week ago @JuhoKokkala raised the following Meta question: Ambiguous [combination] tag -- and the consesus was that [combination] tag should be entirely removed and replaced by [combinatorics] where necessary. (Update: This has been done by now.)

There is a conceptually related tag (261 threads) which is mostly used to refer to permutation-based hypothesis tests, but sometimes refers to combinatoric problems related to permutations. I don't think this lumping is helpful.

Proposed step-by-step solution

  1. Create tag and tag a couple of most important related threads with it.

  2. Go through all [permutation] threads and remove the tag from all non-permutation-test-related questions, replacing with where necessary.

  3. After that, merge [permutation] into [permutation-test]. (This will affect at least 150 threads, so it's better to do it automatically via the merge).

  4. Delete the [permutation] tag as confusing (i.e. delete the synonym mapping).

Update (Feb 22): All done.

$\endgroup$
12
  • $\begingroup$ (+1) About the proposed way forward: I would propose editing [permutation]->[permutation-test] manually in step 4 (instead of leaving them as is and then merging). This is not much extra work since all [permutation] threads need to be checked anyway to see what they are about. Furthermore, this way we would know that the process of checking all [permutation] threads is over when none of them exist anymore. Otherwise, wouldn't this require either some extra bookkeeping system or someone being responsible for going through all the questions? $\endgroup$
    – Juho Kokkala
    Aug 22 '16 at 13:54
  • 6
    $\begingroup$ @Juho It's indeed not much extra work, but it's 150-200 questions pushed to the front page... I can look through 50 questions in half an hour, i.e. in one evening, but everybody will hate me if I push 50 questions to the front page in one evening. $\endgroup$
    – amoeba
    Aug 22 '16 at 13:56
  • $\begingroup$ I did mean that somebody would be personally responsible for going through all the questions, yes. $\endgroup$
    – amoeba
    Aug 22 '16 at 13:58
  • $\begingroup$ Ok, I retract my alternative proposal. $\endgroup$
    – Juho Kokkala
    Aug 22 '16 at 14:01
  • 5
    $\begingroup$ "Hate" seems rather strong. Grimace menacingly at the screen seems more likely. $\endgroup$
    – gung - Reinstate Monica
    Aug 22 '16 at 14:58
  • $\begingroup$ @gung By the way, sometimes I think we are being too strict here. If there is a good retagging suggestion, discussed & approved on Meta, and if somebody has the energy to work through 50-100 threads in one sitting, then why not. I mean, the disadvantages are obvious, but perhaps the advantages outweigh them. $\endgroup$
    – amoeba
    Aug 22 '16 at 15:15
  • $\begingroup$ What would the advantages of mass-retagging-and-bump-flooding-the-landing-page be, compared to slower retagging over a period of days or weeks? (+1, anyway - this exact issue is one I have been wanting to dig into.) $\endgroup$
    – Stephan Kolassa
    Aug 25 '16 at 18:59
  • $\begingroup$ @Stephan "Mass retagging" via a tag merge happens without flooding the front page; when tags are merged, all threads get retagged "silently" without bumping. That is the advantage. Do you want to help taking care of the permutation tag? We could somehow coordinate it between us. It's actually not one of the tags I follow or particularly care about, so I am only happy if somebody else takes charge. $\endgroup$
    – amoeba
    Aug 25 '16 at 21:16
  • $\begingroup$ Confusing tags should not be removed. If they are, new users will create and populate them again. When I work on un-confusing whatever, I first edit [confusing-tag] to say in the excerpt, "This tag is ambiguous. Please use [blah] when asking about this, and [bluh] when asking about that". $\endgroup$
    – StasK
    Aug 26 '16 at 14:01
  • $\begingroup$ @StasK Not sure I understand. Once your [confusing-tag] has 0 threads it will be automatically deleted within 24 hours. $\endgroup$
    – amoeba
    Aug 26 '16 at 14:12
  • $\begingroup$ @amoeba My point exactly. If it is deleted, the text about this tag being confusing is gone, too. So at least one question would have to say for the tag to survive and display the wiki as being confusing. May be I am missing something about how the site operates though. $\endgroup$
    – StasK
    Aug 26 '16 at 14:30
  • 3
    $\begingroup$ @StasK Keeping one old question seems to me to be a really ugly solution. One solution is to get the tag blacklisted, this can be done by our mods asking SE admins. But usually if confusing tag [bla] is de-confused into [bla-foo] and [bla-bar] then nobody will use [bla] anymore because typing "bla" in the tag field will make bla-foo and bla-bar pop up as suggestions. $\endgroup$
    – amoeba
    Aug 26 '16 at 14:41
3
$\begingroup$

There are now just over 200 left which I think could all without loss of information be re-tagged [permutation-test]. Some of the others were [combinatorics], some had little to do with permutation or permutation tests, and one seemed more about entropy so I re-tagged it [information-theory]. Just to clarify as suggested by amoeba in a comment all the rest are permutation test related.

I do not know how to merge the tag with [permutation-test] and I do not have the right magic powers anyway so if a moderator would like to do that it would eliminate the tag.

Update by @amoeba, Feb 21: I have looked through all the remaining [permutation] threads and re-tagged those that I thought were not about permutation tests. There are currently 166 threads left in [permutation] and they are all about permutation tests. I believe we can now go ahead with moving them all to .

Update, Feb 22: Done by @Scortchi.

$\endgroup$
16
  • $\begingroup$ Just to clarify: you are saying that you have looked through all [permutation] questions and retagged all of them (around 60 from around 260 initially) that were not about permutation tests, is that right? $\endgroup$
    – amoeba
    Nov 17 '16 at 21:50
  • 1
    $\begingroup$ That is correct. All that are left are, I believe, about permutation tests. $\endgroup$
    – mdewey
    Nov 17 '16 at 22:05
  • $\begingroup$ I see. You might want to edit your answer so that it says so explicitly. In any case, only moderators have power to move [permutation] questions into [permutation-test] (by declaring it a synonym, merging it, and then deleting the synonym mapping; or something like that). Unless one of the mods comments here in a day or two, I will ping some of them in chat to draw their attention. $\endgroup$
    – amoeba
    Nov 17 '16 at 22:17
  • $\begingroup$ @amoeba edited to clarify as you suggested. $\endgroup$
    – mdewey
    Nov 18 '16 at 9:12
  • 1
    $\begingroup$ Note that despite the tag saying do not use this people are still using it. I just caught one a few moments ago $\endgroup$
    – mdewey
    Nov 18 '16 at 17:21
  • $\begingroup$ I took a brief look at the remaining [permutation] questions and I think I see some that are not about permutation tests. E.g. the first and the fourth newest ones: stats.stackexchange.com/questions/243522 and stats.stackexchange.com/questions/226583. Most seem to be about permutation tests though. I think this tag requires maybe one more look through to double-check before tag is renamed/merged. I might be able to do it, but not right now. $\endgroup$
    – amoeba
    Nov 18 '16 at 21:25
  • $\begingroup$ @amoeba I took the view that if they said they were doing a test and it involved permutation in some way then it was a permutation test even if it was not the classical permutation test. I would be delighted if someone checked my decisions of course. $\endgroup$
    – mdewey
    Nov 18 '16 at 22:04
  • $\begingroup$ @amoeba I see that people are still using the tag despite the warning message so perhaps I need to have another go. Perhaps I should manually rename the ones where I am absolutely sure so there are fewer for someone else to check. $\endgroup$
    – mdewey
    Dec 2 '16 at 14:43
  • $\begingroup$ It took me some time, but I have started looking at it. Hopefully we will manage to close this issue soon. $\endgroup$
    – amoeba
    Feb 13 '17 at 11:01
  • $\begingroup$ @amoeba OK I will try re-tagging some of the ones where I am 100% sure it is a permutation test and see what we have left. $\endgroup$
    – mdewey
    Feb 13 '17 at 13:49
  • $\begingroup$ That's not what I meant! If it's the majority of 200 Qs then I'd say it doesn't make sense. I was rather going to look through the threads and retag those that are not about permutation testing. I know you did the first pass, I can do the second, and hopefully this should be enough. $\endgroup$
    – amoeba
    Feb 13 '17 at 13:57
  • $\begingroup$ @amoeba OK, you are right of course. $\endgroup$
    – mdewey
    Feb 13 '17 at 14:03
  • $\begingroup$ If you want, you can take a look at the [randomization] tag and try to figure it what to do with it... It does not have a wiki excerpt, is probably ambiguous, and many Qs are about permutation test aka randomization test. I suspect we need randomization-test tag mapped to permutation-test, and also randomized-experiment tag or something similar. $\endgroup$
    – amoeba
    Feb 13 '17 at 14:15
  • 1
    $\begingroup$ Just tag them with [permutation-test]. Some sources make a distinction between "permutation test" and "randomization test" (whether all possible permutations are used or only a subset of them), but IMHO it's irrelevant: in practice everybody uses only a subset of possible permutations and both names are used for that without distinction. We will make a synonym later. This was discussed in chat some time ago and nobody protested. $\endgroup$
    – amoeba
    Feb 16 '17 at 16:02
  • 1
    $\begingroup$ I accepted your answer because this has been implemented! $\endgroup$
    – amoeba
    Feb 22 '17 at 14:03

You must log in to answer this question.

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