For discussion of some of the tag proposals a distinction should be drawn between:
A. objective attributes of questions derived solely from the content of the question. Examples: is the question over 500 words in length? Does it give more than zero sources for the material posted? (These could be instantiated as hypothetical tags [long-question] or [contains-sources], which I am not proposing but just providing as examples of objective and presumably reasonable question descriptors that go beyond the mathematical subject tags.)
B. "meta-tags" or "dependent tags" as defined in the Death-of-Metatags policy posting on Stackoverflow. These are subjective or questioner-dependent attributes of the question: is it an [easy-question], is it [homework] (where not reported in the question), etc. Such tags are (in the DeathOf posting) said to be ambiguous or contentious and are discouraged for various reasons.
I believe that tags of type A are unobjectionable and in some cases highly desirable as a method of content filtering. For example, some people might be very interested to know, before clicking on a question, whether it is over 500 words long or contains sources, just as they might be interested in the number of upvotes, answers or views. Generally, more information is better, and the information one would ideally like about a question is not limited to the mathematical subject tags.
There seem to be posters here who want math-subject-tags only. If so, it would be good to hear their arguments for why number of words or other such indicators should be off-limits for tags.