# Do we want tag warnings, and for which tags?

A new feature has been introduced this summer, tag warnings.

These warnings are meant to provide users with just-in-time information to help them avoid common mistakes in the tag they’re using (e.g. SQL questions should mention the engine they use).

Here’s what they look like:

(More information on this here, from which the above text and image were taken.)

It seems as if it could be a really helpful feature, especially for new users, so they can avoid common pitfalls. Indeed most people seem not to read tag excerpts, but they might read a yellow warning popup. A few examples of tags for which it could be helpful are , and its friends, , (as opposed to ) and similar cases, , extremely general tags like , etc. See also this related discussion about "off-topic tags".

I suggest that we follow the format laid out at meta.academia.SE: write down the tag name, the proposed warning text in a block quote, and some context.

• Mildly relevant: meta.math.stackexchange.com/questions/11140/… – Asaf Karagila Nov 5 '15 at 11:11
• We have a list of comment templates for tags that are often used incorrectly. (However, the list is almost surely uncomplete.) This might help with identifying some suspects. – Martin Sleziak Nov 5 '15 at 11:30
• More than warnings, why not prevent people from posting questions with certain tags without them being accompanied by some other appropriate tags, like here on Meta MSE? For example, (analysis) alone shouldn't be allowed, same with (abstract-algebra) and (discrete-mathematics). – Git Gud Nov 5 '15 at 11:34
• This is so rude. :-) – quid Nov 5 '15 at 12:14
• @quid: But is it hacking the system? – user642796 Nov 5 '15 at 13:12
• @GitGud There isn't built-in support for "if you use this tag, put another tag". And even if there was, I'm afraid it would be more of discrete-mathematics + proof-whatever-autocomplete-gives. – user147263 Nov 5 '15 at 14:04

If your question is not about the area of mathematic logic called proof theory, do not use the tag. Questions requesting proofs for a particular statement, or about how/why a particular proof of a statement works, should be tagged with the mathematical subject the statement lies in.

This tag is almost universally used incorrectly as a substitute for the blacklisted intrinsic (similar to and , I imagine). However it names an honest area of mathematics and should not be just done away with.

Someone should probably got through the tag and clean it up. (I guess I'm a someone.)

• Hear hear! I'm tempted to write one for the [set-theory] tag "If you added this just because your question involves sets, please remove the tag." :-) – Asaf Karagila Nov 5 '15 at 11:33
• Chiming in with Asaf. If the tag [division-algebras] is added because the question involves divison of objects that the asker encountered in a (high school) course titled algebra it is just sooo wrong. I have considered removing that tag altogether. There aren't very many questions using it properly, and a significant fraction of new questions with that tag use it incorrectly. I am also pointing fingers at the answerers who do not bother checking the tags even though they should know better. – Jyrki Lahtonen Nov 5 '15 at 22:43
• I am joining @Jyrki in the finger pointing business to those who know they should retag something, have the time to write an answer, but don't have the time or inclination to fix the tags. (Of course I'm not referring to people who on occasion don't retag; I'm talking about those who consistently don't do it.) – Asaf Karagila Nov 6 '15 at 7:08
• @AsafKaragila, I think part of the problem is that "elementary set theory" means "I'm dealing with sets," whereas "set theory" means "I'm dealing with universes of sets," e.g. models of ZFC, or ETCS, or otherwise. Obviously, this is pretty confusing. Too late to change it now, of course, but it would have been nice if Godel and Cohen had said that their work was part of "universe theory" or "cumulative hierarchy theory" or "models of ZFC theory," as opposed to "set theory." – goblin Nov 7 '15 at 4:18
• @goblin I don't think this is really what Asaf is talking about (and is a bit beside the point for this answer, but alas). Not infrequently do users ask a question where sets are present and tag them set-theory, but set-theoretic notions (elementary or otherwise) are not really central to the question. An example is here, where the set-theory tag wasn't removed until after both answers were posted, and one of the answerers made an inconsequential edit to the body. – user642796 Nov 7 '15 at 5:24
• I think I have an easier solution. Contact the guys from proof theory and ask them to come up with a different name for their field. Perhaps "Glorbax theory". Then we can just burninate and blacklist this tag! – Asaf Karagila Nov 9 '15 at 13:40
• This is also an issue with the "logic" tag, which is used a significant number of times to mean "logical thinking is required to do math". It is intended for questions on mathematical logic. – Carl Mummert Nov 12 '15 at 16:23
• @Arthur Fischer: I have cleaned up numerous questions in this tag in the past few days. The majority of them were about elementary number theory (maybe "intro to proofs" courses). Almost none were really about proof theory. Is the tag warning active yet? – Carl Mummert Dec 7 '15 at 15:44

Instead of generic (analysis) tag, use a tag for a specific branch of analysis such as (real-analysis) or (complex-analysis) or (functional-analysis), etc. For data analysis, use (data-analysis).

This was discussed last year: Why is the tag (analysis) useful?

(assuming we want to keep it)

The proof-writing tag is for advice on clear and rigorous presentation of a proof that you have in mind. If you don't have any proof yet, don't use it.

This is my interpretation of what the tag is supposed to be about; the actual usage, of course, is that users want to tag with "proof" (my question is about a proof, right?) and pick one of offered options, such as this one.

• I have often questions like "How to improve the clarity of this proof?" but I don't post them because they will with no doubt down voted or otherwise rejected because they are not "answerable". Or does that depend on who you are and know? – nilo de roock Nov 8 '15 at 17:42
• If you can identify a particular step of a proof that you feel could be presented better, I don't think the question would be treated as "unanswerable". – user147263 Nov 8 '15 at 17:56
• @ndroock1 The way I understand it, this is what proof-verification tag if for. At the moment, the consensus is that such questions are allowed. It is possible that this policy might change in the future, as discussed here. – Martin Sleziak Nov 13 '15 at 12:06

Linear algebra is not algebra with functions such as $y=mx+b$; it works with vector spaces. Make sure your question involves vector spaces before using this tag.

I've seen many new users use in place of .

• This is assuming that the person has used in incorrectly in that specific way... – Ali Caglayan Nov 12 '15 at 7:54
• A user might be dealing with a system of linear equations in some form and not even realize that these can be formalized using vector spaces, both for the space of all variable assignments and (in the homogeneous case) for the space of solutions satisfying the equations. I don't know how likely such a user is to pick the linear-algebra tag, but I think it might be possible. – MvG Nov 13 '15 at 1:09
• This is weirdly phrased. Linear algebra is algebra with functions such as y = mx + b... I understand what you're aiming for here, but it could use rephrasing. Maybe "Linear algebra is a college class. If your question is about equations of lines from high school algebra use algebra-precalculus." – Noah Snyder Nov 20 '15 at 15:05
• Maybe something like "This tag is not for questions about linear functions of real numbers from precalculus/calculus. Linear algebra is about vector spaces, linear maps, matrices..." – Najib Idrissi Nov 20 '15 at 18:10

Note that not every question involving algebra and geometry belongs in the subject of algebraic geometry. See the tag wiki if you are unsure.

The tag wiki explicitely warns against abuse of this tag, and I've seen some incorrect use myself despite that.

• Maybe it's better to directly give a positive definition? – Ortomala Lokni Nov 16 '15 at 14:30
• @OrtomalaLokni: I couldn't think of a short yet precise and understandable definition. I thought about “if you post using this tag, you should know what an algebraic variety is”, but that might discourage valid questions about elliptic curves and the likes. I thought about focusing on “zero-sets of a system of non-linear equations” but feared that this would still leave us with many elementary questions about axis-aligned ellipses and how they intersect with various stuff. If you have a good suggestion let's hear that. – MvG Nov 16 '15 at 15:01
• We can use the official user guidance: "The study of geometric objects defined by polynomial equations, as well as their generalizations: algebraic curves, such as elliptic curves, and more generally algebraic varieties, schemes, etc. Problems under this tag typically involve techniques of abstract algebra or complex-analytic methods. This tag should not be used for elementary problems which involve both algebra and geometry. " But it's maybe too long. – Ortomala Lokni Nov 16 '15 at 15:25
• @OrtomalaLokni The tag excerpt is displayed when a user selects a tag. I guess some redundancy is good, but these warnings are intended for people who do not know what algebraic geometry (in this case) is, and there's very little chance that the algebraic-geometry tag would be a good fit then. – Najib Idrissi Nov 16 '15 at 16:10

There are so many questions using this mark, but which are simple calculations for high school / undergraduate courses (and should use the tag). I've tried to edit these posts, but seeing that the same mistakes are made every day is a little disheartening.

I have recently found a few questions tagged with which were just about the analysis of a specific function (or functions satisfying a certain relation like $f(x+y) = f(x)+f(y)$) and not related to actual functional analysis at all.

I am not certain, however, if the problem is big enough to warrant an extra tag warning.

• I see those, but I think a warning would be an overreaction. Tags generally get used haphazardly based on partial word matches. [functional-analysis] is in pretty good shape. – user147263 Nov 11 '15 at 21:47

This tag is for ordinary differential equations. If your question is about partial differential equations use the tag instead.

This is a common one and a simple warning as such would clear much confusion about the use of this tag especially for new users.

• Could we have a full stop and a new sentence before that if? – TRiG Nov 12 '15 at 11:11
• A possible issue is that users don't necessarily know the terms "ordinary" and "partial" if they are in a course on ordinary differential equations, they only know they're taking a course of Diffy Qs, whatever that means. – user147263 Nov 13 '15 at 2:57
• @NormalHuman That is a non-issue as this only applies to questions about partial differential equations that are not about ODEs. If a user does not know what an ODE or a PDE is then they are most likely asking a question about an ODE anyway. – Ali Caglayan Nov 13 '15 at 17:50
• Why is the tag called "differential-equations", if it only covers ordinary differential equations? Shouldn't it be renamed to ordinary-differential-equations? (Or maybe a synonym ordinary-differential-equations should be created, and should be made the canonical name of the tag?) – fonini May 9 '16 at 2:31
• @fonini I believe this is because the name will not fit. – Ali Caglayan May 9 '16 at 14:11

Half the tags are math questions unrelated to mathematical pedagogy. There are also a lot of questions that are "seeking personal advice" like what class/book to take/use next.

I would advocate a warning for (or a name change) the tag complex geometry. This is supposed to refer to complex manifolds and complex algebraic geometry, but for a while the bulk of the posts were elementary questions about the complex plane such as one might find in precalculus or beginning undergraduate complex analysis. I suspect that changing the tag to complex differential geometry or complex manifolds might help with this, although each of those is a bit more restrictive than I'd like.

(Mixing roughly the structure of arjafi's text for and the tag excerpt for )

If your question is about projections in linear algebra (for example, projection matrices), please use the tag. Projective spaces are a subject of geometry that deals with spaces which can be seen as the set of lines through the origin in some other vector space. You can also consider the tags and , if you think they fit your question.

• fonini: I think that in addition to (linear-algebra), also the tag (linear-transformations) might be useful for questions about projections, (matrices) for projection matrices. – Martin Sleziak May 9 '16 at 5:54
• Done. Shouldn't this be CW? – fonini May 9 '16 at 16:56