Some spontaneous thoughts on tagging in general, inspired by a thread not fully on the subject thus recorded here:

I would consider it as a good idea if there were more organization in the tagging system and if there was something like a "tag-tree" on meta or somewhere.

I tried to work a tiny bit in this direction via mentioning explictly in some tag-wikis that I wrote both more general and more specialised tags that could be sued in addition or instead of some tag.

One major obstacle against a more systematic tagging system is that there are so many tags already. I argued against this already a long time ago, but some held the opinion that very specialised tags served some purpose.

(The discussion than was not very detailed so I did not really understand the reasons, but my personal impression/guess is that some people see the prupose of tags differently than I. Namely, rather as flashy keywords than as meta information.)

Everbody who intends to put some more structure in the tagging system can expect my abstract and concrete support. Only, I think it is important to try to do it in such a way that it still works somehow for people not understanding the system. And, also to move only slowly.

Another issue with reorganisations is that some (and I do understand this, somewhat, but only somewhat) are very averse against lot of traffic on the frontpage. So that things only can proceed very slowly. Sometimes this is frustrating. In a spare half-hour here or there one could do a lot of tagging if one has a clear plan. But one must not. But do three, and wait for a day.

@SébastienPalcoux Hi and welcome! Thanks for stopping by in this room.

@quid: That is interesting. Nevertheless, I am confused about how those things could work.

First of all, the distinction between keywords and meta information is unclear to me.

Second, the idea of a tag tree of some sort is highly attractive, and could be very useful at least as a guideline. However, if one were to try to implement such a system more rigidly (say, perhaps to replace the current system), I do not see exactly how it would look or function, or how usable it would be. Can you give a hypothetical example?

Third, I agree that it would be exceedingly useful to add more structure of any kind to the tag system. However, one has to also remember that (barring impossible perfect design) the structure will be inevitably subverted by the users of mathoverflow.

Just to give a very rough example, consider that the top level tags currently do not even cover all questions which appear (e.g. linear algebra questions). Also, in a complex system, trees usually have cycles (thus making them not trees :) ).

Hi @quid. Here is an observation about editing: even when we are careful to write a good post, there are often many comments suggesting to improve it. But in the same time, people are unhappy with multi-edit. The fact we can't edit as freely as we want is a failing of MO, in my opinion. Perhaps this point I'm talking about affects only the rough recent users, as me...

@Sébastien Palcoux: Unfortunately, I have very little to say about that. I, too, would like stuff to be a bit more editable. Other than that, I would like to direct your attention to the feature request "Minor edits, subject to review" started by Scott Morrison, just in case you were not aware of it. Such a feature could go a long way. Needless to say, this feature request is unlikely to go anywhere...

Also, welcome to this lounge, @Sébastien!

By the way, @quid, please do not mistake my comments above as significant opposition on my part. I just like to play devil's advocate.