A solution might be to extract manually the technical usages in other mainly to be created tags, and to remove the rest. (I would be willing to contribute to this 'extraction' effort.)

On a different subject (continuing from a couple comments above): "analytic" is meanwhile/at the moment ready for merging into "analytic-number-theory" (at the moment 6 questions, which I add on the small chance that something appears in the interim.)

Are there any objections to this? Or other/better suggestions? (The amount of manual retagging is not large but nonminimal, say 5+eps. I will not do all at the same time.)

And, at the moment the five question tagged differential could be merged into dg.differential-geometry (I just extracted one where the also existing differentials seemed needed).

mathematical-physics -> mp.mathematical-physics

quantum-group -> quantum-groups

computabilty -> computability-theory

topos->topoi

ha.homological-algebra -> homological-algebra

refe -> reference-request (or tag-removed as the question is already tagged reference request)

graph.theory -> graph-theory

Some more complicated ones:

derived-categories got created the 'wrong' singular exists with about 100 questions. (This could be a good tactic for getting a taxonimist tag; search a singular tag with a lot of use and create the plural ;D)

Not sure what to do with

examples-counterexamples

AFAIK this is math.SE practise, we have both separately; the current one asks for examples and I am under the impression OP only used it out of math.SE habit. (I once had a conversation on a similar instance where this was given as reason.)

In fact, I just noticed we have

counter-examples and counterexamples (and examples)

as tag.

One possibly controversial one:

Recently non-associative-algebras and nonassociative-algebra got created (by different users).

The later usage is more like algebra as field and not structure. So, in principle, one could say there is a difference; but due to the general scarcity my personal opinion would be that perhaps one of the two would suffice. Likely i'd preserve the latter.

Thanks in advance for merges!

There is an (established) tag 'continuum-hypothesis'. Somewhat recently a tag 'gch' got created in addition (it has now 2 questions tagged like so, the second by me, when an old question became reactivated.) Yet, just know a question got asled that seems directly related to the generalized continuum hypothesis but is still tagged only continuum-hypothesis.

In my mind this raises the question whether it is useful to have both tags, for general "size" considerations but even more so since this latest question reinforced my general conviction that consitent usage of fine-grained tags simply is unlikely to happen.

Could the pluralized 'continuum-hypotheses' as common tag makes sense?

Singular/plural duplicates (I only give the singular to be merged into the plural):

derived-functor

finite-group

galois-representation

graph-coloring

line-bundle

locale

perfect-matching

quiver

random-graph

semiring

simplicial-complex

Other 'duplicates' or clear 'errors', sometimes I am not sure which form is better

nonstandard-analysis <-> non-standard-analysis

preduals <-> pre-duals

char-p -> characteristic-p ; in addition there is also positive-chracteristic (perhaps in fact both former ones should go into this one)

reference -> reference-request

refer -> reference-request

ha.harmonc-analysis -> harmonic-analysis (no arviv cat.)

geometric-topology -> gt.geometric-topology

Near duplicates (perhaps some are debatable)

dgalgebra -> dg-algebras (at least I believe this makes sense)

derived -> derived-algebraic-geometry (there is only one use of derived)

height-functions -> heights (or the other way round)

algebraic-numbers -> algebraic-number-theory (the former is only used once while the tag is old)

convex -> convexity (the convex tag is somewhat harmful as it "distracts" from convex-something tags; and I think this makes sense for all tagged like so at the moment; there was one where convex-optimization was better I retagged that one manually)

inner-model-theory <-> inner-models (perhaps?)

harmonic -> harmonic-functions (the latter does not exist, but I think it fits for all current uses while I am slightly worried to see 'harmonic numbers' for example making 'harmonic' a strange mix; there is perhaps one exception a question on hermonic polynomials, but since these are AFAIK also harmonic functions it is not so bad)

computability -> computability-theory

real -> real-analysis (seems to make sense at the moment, and real is perhaps a tag to be eliminated)

Added

A new singular/plural

arithmetic-group -> arithmetic-groups (the correcter plural is new)

And another arXiv without prefix

differential-geometry -> dg.differential-geometry (11 questions)

One more close duplicate

hyperbolic -> hyperbolic-geometry (the tag hyperbolic was somewhat recently born as a broken hyperbolic-geometry; this resurfaced on the front page, which I thus fixed right away, another usage of it was for hyperbolic pde, I created that tag now and retagged one; the remaining is again hyperbolic-geometry)

Several arXiv-like tags, mistyped, or just missing the prefix or otherwise extremely close, and would need merging into the arxiv one:

statistics -> st.statistics

gt.general-topology -> gn.general-topology (yet not gt.geometric topolgy :-) )

ag.algebraic-topology -> at.algebraic-topology

k-theory -> kt.k-theory-homology (this might be debatable, but still seems to make sense, IMO)

functional -> fa.functional-analysis (at the moment there seems to be no single usage of this tag in a truly meaningful way, yet on the other hand it is sometimes abused as building block of "functional analysis"; if ever there is a need the plural "functionals" could be created; yet now it would seem renaming is a lot worse than just merging)

category -> ct.category-theory (at the moment there is just one tagged like so, but I do not want to bump that one)

control-theory -> oc.optimization-control (same as above)

number-theory -> nt.number-theory

numerical-analysis -> na.numerical-analysis (I mention the last two again, as due to my sub-optimal way of reporting them earlier I believe they were overlooked)

Some singular/plural issues (where both exist):

curve -> curves

journal -> journals

manifold -> manifolds

norm -> norms

tensor-product -> tensor-products

Some other things

linear -> linear-algebra (I checked this makes sense at the moment; in some cases manual changes would be "cleaner" as they would not leave tangling 'algebra' tags but 'algebra' seems "lost" as a tag anyway so it might not be worth the bumps)

complexity -> computational-complexity (also makes sense at the moment)

community-wiki -> tag-removed (here I am not sure, but to have three out of many cw-questions tagged cw seems pointless yet could have the bad side-effect that some new users might think tagging like so is what "making community-wiki" means)

However, there are some instances where both singular and plural exist, yet the singular is frequent and the plural one very rare (indeed unique!). To have both seems quite unfortunate, yet I am hesitant to simply 'kill' the more correct though infrequent ones.

The examples I am presently aware of are:

hilbert-space(s) 93 vs 1 unitary-representation(s) 27 vs 1 blow-up(s) 37 vs 1

So this is either a request for the merges singular to plural or a request for permission for me 'killing' the plural ones.

Added: analogous problem for "number-theory" (seven questions)

Added 2: likewise for "numerical-analysis"

In addition there is a single one tagged ho.historical-overview that also should be merged (since it is only one I could do so myself but the question is very old, and since I am already mentioning history tags I thought I leave it for the moment)

Thanks in advance!

matrix -> matrices

combinatorics -> co.combinatorics

Some tags get created by accidentally splitting hyphenated tags (e.g. [algebraic] [topology] instead of [at.algebraic-topology]). These need to be sorted out periodically to avoid the proliferation of uninformative tags. Major problems regularly occur with the following:

• algebra
• algebraic
• analysis
• differential
• equations
• geometry
• geometric
• theory
• topology

It would be helpful if the community participated with these corrections. If you happen to notice one of these tags on the front page, please correct it immediately. If you happen to be in the mood to go through the lists and sort them out, please send a list of corrections to moderators@mathoverflow.net rather than bumping tens of oldies to the front page.

I do not know of a way to prevent the creation of tags. As long as we are talking about things we'd like to see, perhaps there should be a system (outside our brains) that not only included tags, but encoded relationships between them together with more specific information concerning their pertinence to the questions.

If both have the same amount, just pick one and rename it.

Either way we are bound to cut the retag by at least a half.

I realized that there's only 17 questions tagged [presentation], so probably the fastest solution is indeed retagging them one by one. They are so few, that I could even volunteer for the job. :)

1. Is there a smarter way than retagging them one by one?

2. how should we rename the two tags (or at least one of them)?

• algebra
• algebraic
• analysis
• differential
• equations
• geometry
• geometric
• theory
• topology

It would be helpful if the community participated with these corrections. If you happen to notice one of these tags on the front page, please correct it immediately. If you happen to be in the mood to go through the lists and sort them out, please send a list of corrections to moderators@mathoverflow.net rather than bumping tens of oldies to the front page.

Scott, since it has no information content, we should merge it into [tag-removed] and let time sort things out...

http://mathoverflow.net/questions/tagged/sequences-and-series

http://mathoverflow.net/questions/tagged/sequences

the latter is perhaps redundant?

[gn.general-topology] and [topology] and [general-topology] should all be merged into one tag, the rest should be made synonyms.

• http://mathoverflow.net/questions/tagged/differential-graded-algeb
• http://mathoverflow.net/questions/tagged/differential-galois-theor
• http://mathoverflow.net/questions/tagged/differential-graded-lie-a (do we really need a tag for differential graded lie algebra?)
• http://mathoverflow.net/questions/tagged/pseudo-differential-opera (I'm not entirely convinced that we need a tag for pseudo-differential operators. But if we do, it maybe better to name it pseudo-DOs or pseudo-diff-ops (too bad we can't use greek letters in tag names, else \psi-DOs would be standard).)
• http://mathoverflow.net/questions/tagged/combinatorial-number-theo
• http://mathoverflow.net/questions/tagged/combinatorial-hopf-algebr
• http://mathoverflow.net/questions/tagged/symmetric-monoidal-catego
• http://mathoverflow.net/questions/tagged/polya-redfield-enumeratio
• http://mathoverflow.net/questions/tagged/multiplicative-number-the
• http://mathoverflow.net/questions/tagged/computational-number-theo

Also, I now challenge you to also fit "harmonic analysis" in that tag name (it technically also belongs, you know). :-p

The change is done.