When coming to post a question, one looks for the most fitting tags.

After a while, it may be apparent that a new tag is in order. Good examples are the and and so on.

However, sometimes when the topics are of higher level, it might be as apparent (i.e. there is not a substantial amount of questions, and it is doubtful that such amount will ever exist at all).

What should be a good rule of thumb for "when to add a new tag"?

Edit I'm not posting this question in order to advocate against new tags, I'm sincerely wondering when it is appropriate to split an advanced topic into a new tag, especially when knowing there will not be more than a dozen or two questions per year.

Edit 2: (TB) I'm adding to this thread in order to revive it and in order to avoid creating a (possible) duplicate. In the last few hours there were the following new tags:

How should one deal with them? In my opinion they are all way too specific and in order to make them useful one would have to go through many questions that should belong to them and add the tags. Should one delete them unilaterally (i.e. remove them from the question and wait for the engine to burn them) or create a new answer on the Tag merging and synonyms thread?

  • $\begingroup$ @Theo: Not really, I cannot judge about that. I can only say that the name is not to my liking :-) $\endgroup$ – Asaf Karagila Mod Jun 28 '11 at 22:01
  • $\begingroup$ It's hard to formulate some general rule... But probably it was discussed elsewhere (on MSO, meta.MO,...) already? Does anybody know relevant link? $\endgroup$ – Grigory M Jun 29 '11 at 20:21
  • 1
    $\begingroup$ @Grigory: Dear Grigory, Here are some relevant links from meta.MO: tea.mathoverflow.net/discussion/34 tea.mathoverflow.net/discussion/1071 Regards, $\endgroup$ – Matt E Jun 29 '11 at 20:54
  • $\begingroup$ @Grigory: I was not looking for a "rule", but rather a "rule of thumb". How big is the lashing stick before it is too big and needs a new tag. In particular I am asking for [set-theory] sub-tags which are not abundant. While I like it like that, I also feel that there might be a slight benefit to it sometimes. $\endgroup$ – Asaf Karagila Mod Jun 30 '11 at 6:43
  • 2
    $\begingroup$ @Asaf I like a rule of thumb from (AFAIR) Qiaochu Yuan: if you can't imagine a user adding a tag to favorite/ignored, then it's a bad tag; and if you can -- probably (not always, but probably), it's OK. $\endgroup$ – Grigory M Jun 30 '11 at 7:57
  • 1
    $\begingroup$ Asaf: I added a follow-up to your question since I wasn't sure if it was worth opening up a new thread for that. I hope that's fine with you. $\endgroup$ – t.b. Aug 22 '11 at 8:39
  • 1
    $\begingroup$ My humble opinion: automata looks fine for me; having to specify that you're thinking about the pushdown sort in the tag looks way too specific... $\endgroup$ – J. M. ain't a mathematician Aug 22 '11 at 8:50
  • $\begingroup$ @Theo: It's fine. $\endgroup$ – Asaf Karagila Mod Aug 22 '11 at 8:50
  • 2
    $\begingroup$ @J.M. I agree with that. A further qualm I have with (context-free-grammars) and (pushdown-automata) is that they are just two ways of looking at the same thing. A similar remark would apply to (regular-languages) and (finite-state-automata) (the latter doesn't exist, yet). See Chomsky-Hierarchy $\endgroup$ – t.b. Aug 22 '11 at 8:56

(Answer to TB's edit)

  • If you are not sure whether a tag belongs, bring it to Meta.
  • If you feel strongly that the tag does not belong, delete it unilaterally.
    • But don't get into a retagging war: if the OP disagrees, bring it to Meta and flag for Moderator attention.

Personally I think all the new ones mentioned by TB, with the possible exception of "group extensions", are too specific. I note (without further comment) that four of the tags there were all created by the same user. I've alerted that user to this discussion.

  • 1
    $\begingroup$ Re: last paragraph. Yes, maybe I should have done this myself. I didn't because I didn't want to stir up more dust than necessary and it is more of a coincidence that these are all due to that user. See also the comments here. $\endgroup$ – t.b. Aug 22 '11 at 14:39

You must log in to answer this question.

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