The tag system is fine, and the names/explications of the tags clear enough. But why is there no way to add facultatively primary (and secondary?) AMS classification numbers, like in many journals?

I don't mean replacing the tags, but as the order in which the tags appear for a given question is rather random, the primary AMS classification might especially be useful for sorting or searching purposes. The majority of tags in the thousands of already existing questions could be automatically converted into corresponding AMS numbers, though that would not give a hint about which one should be considered primary for a given question.

  • 1
    $\begingroup$ The tags do not appear in a random order (they are sorted by size of the tag, at least the first is the largest). Except for old question (before upgarde) where the order is as specified by user. It is true there is now no notion of a *primary*/first tag (and there never really was one), but I do not consider this as a big deal. $\endgroup$
    – user9072
    Commented Jan 30, 2014 at 21:35
  • 2
    $\begingroup$ The answer to why there is no way is simly because the softwarre is used for many different things and for many/most there is not such thing as a standardized subject clssification. One might ask for this, but then, we have the toplevel tags from arXiv with the idea of serving some vaguely related purpose. Already this adds enough complexity to lead to somewhat constant confusion. I am really not to optimistic any change will lead to an improvement in pratice. By contrast everything that adds complexity will result in more confusion, inconsistencies and generally worse situation. $\endgroup$
    – user9072
    Commented Jan 30, 2014 at 21:38

1 Answer 1


When MathOverflow was created, the site was seeded with a few initial tags, one for each arXiv category which are distinguished by starting with a two-letter acronym:

Because many of these have attracted considerable following over the life of MathOverflow these remain the major categorization tags for the site. As a rule of thumb, every question should try to use one of these tags along with more specific tags.

The reason to use this categorization instead of MSC is mostly historical, but the basic reason why the arXiv chose this categorization also applies here:

Why not the Math Subject Classification?

The categories (or subject classes) of the xxx math archive bear some resemblance to the top-level sections of the Math Subject Classification (MSC) used by Math Reviews and Zentrallblatt für Mathematik. The two classifications, however, differ in their intent and history, and the math archive advisory committee decided that strictly following the MSC was unworkable. For example, Quantum Algebra, the most popular category in the math archive, has no counterpart in the MSC. It would be impractical, perhaps even counterproductive, to reorganize this category into MSC sections. Each category of the math archive must maintain a self-sustaining stream of new submissions in order to be useful; the committee decided that the MSC sections are in places too specific or too dated for this purpose.

In fact, if we were to reorganize the MathOverflow tags, we would definitely prefer a classification where each major tag gets a steady flow of questions than a well-organized taxonomy like the MSC.


You must log in to answer this question.

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