How should our tags look like in Q&A?

Question

How should our tag names look like in Q&A?

On the blog, it was suggested that we could use ArXiV tags. I had upvoted the suggestion then, but now that I think about it, I don't like the idea very much. Not in Q&A. ArXiV tags are quite broad, like gr-qc, etc. There is no way it can have a (relatively) more specific tag like "heterotic-string", or "horava-witten-domain-wall", or something like that. They are also not particularly helpful for our SEO. They may seem to make the site look more professional, but they don't. They look funny, and make people roll their eyes, thinking "What a pathetic attempt to look professional or high-level.", effectively chasing them away from the site. 

It has however been suggested on Meta.TP.SE that while the tagging system need not be completely ArXiVised, every post must have (not technically must, but in terms of rules) an ArXiV tag. However, I agree with Piotr Migdal's answer there. I think that ArXiV tags are also skewed against Experimental Physics and Phenomenology. It also excludes a lot of fields, especially in Experimental Physics and Phenomenlogy.   

Another option is PACS, which has been suggested in Meta.TP.SE. However, this is too cryptic, and is not very helpful for our SEO either. Who in the world will understand what physics field 12.10.Kt is about without checking?   

Another possibility is MathsOverflow style top-level tags, like mp.mathematical-physics, or something like that. But Maths Overflow uses it as a replacement for a categorial system, which we already have (Note that as a mathematical site, ag.algebraic-geometry or mp.mathematical-physics make sense). We could implement the same concept for more specific things, like st.string-theory, qft.quantum-field-theory, lqg.loop-quantum-gravity, or something. But I think that my comment about "What a pathetic attempt to look professional or high-level." still applies (note that this does not apply to MathsOverflow because it uses such a structure as a replacement for a categorial system, which means that there is a practical reason.). 

Or, we could use the existing tag system, where tags arise out of need, and are physics terminology in lowercase with hyphens instead of spacebars. This makes most sense, I think. Why, even research papers use such tags (keywords). 

Answer 1

Maybe the ArXiv tags can be used to organize questions in the Reviews section, such that the ArXiv categories are the first level, the second level consists of the ArXiv tags, the keywords provide the third level, which can be specialized further by more specific physics terms at the forth level and below etc ...

In the Q&A part, I think it is enough to use tags as they arise out of need.

Comments

Agreed. Regarding the reviewing, I agree but note that many ArXiV papers don't have keywords on them, especially those not submitted to a journal, so we will have to do that ourselves during the retagging. 

Also note, that the easiest way to have this "hierarchial tag system" for ArXiV papers is through sub-sub-categories, so the actual tags will have to be treated separately.