3
$\begingroup$

Consider tags such as , , , , , , , , and possibly others I missed.

Do we really need all these tags about basic mathematical topics? I feel like there's not much point in having them, as questions whose main topic is something like "inner product" is either better asked on mathematics, or it's about basic topics and sufficiently covered by and maybe .

$\endgroup$
9
  • $\begingroup$ Note that [hilbert-space] is even a tag on physics, which otherwise (I think) doesn't really have math concept specific tags. -- More generally, I feel that there might be tags which are sufficiently QI-specific to justify their existence here, such as [tensor-product], which are concepts much more common in QI than in most other areas. $\endgroup$ Sep 6, 2023 at 18:56
  • $\begingroup$ @NorbertSchuch I'm not questioning that these topics are important in QI; they clearly are. But that doesn't mean the tags themselves are useful. Eg what's the point of a tag such as tensor-product? Tensor products are involved in one way or another in lots of questions, but we likely don't want to tag them all with it. And if a question's primary topic is indeed tensor products, I'd question how on-topic it would be. It would probably be either very basic, so a textbook-and-exercise question, or advanced enough to be better suited for math.SE. What would be an actual use case for it? $\endgroup$
    – glS Mod
    Sep 6, 2023 at 22:05
  • $\begingroup$ Fair point. Admittedly, the "hilbert-space" on physics might be more appropriate there since it only applies to questions relating to quantum mechanics (and it might well be replaced by the latter tag there), while here all questions deal with Hilbert spaces in one way or the other. $\endgroup$ Sep 7, 2023 at 18:37
  • 1
    $\begingroup$ Mathematics-related tags are a clutter (this is probably true more generally, but I'll stay on topic). It seems to me that a big part of the issue is that they are too fine-grained. However, coalescing them all into mathematics seems too coarse and unlikely to be helpful. It might be more useful to tag math-related questions with a subfield, e.g. linear-algebra, group-theory, functional-analysis, number-theory etc. $\endgroup$ Sep 8, 2023 at 4:11
  • $\begingroup$ @AdamZalcman I agree that that could be a good compromise. Have tags for general mathematical topics, and remove all other topics about specific basic mathematical topics (90%+ of which are about elementary linear algebra anyway) $\endgroup$
    – glS Mod
    Sep 8, 2023 at 12:29
  • $\begingroup$ Wow, I just realize there is an amazing number of rather specialized tag. (And amazingly, someone even wrote short texts for most of them.) @glS Is there a way to suggest to map existing tags onto another? (Random example: qiskit-primitives should probably map to qiskit -- is there a way for me to suggest this?) $\endgroup$ Sep 8, 2023 at 17:43
  • 1
    $\begingroup$ @NorbertSchuch yes, there's a lot of tags that should be fixed. Some of them I've always thought it was weird to have them both, but was never entirely sure if there were subtleties I don't know about (the various ibm-x come to mind: do we really need all of them?). The easiest way to implement "mappings" like you're suggesting is to use synonyms. You can suggest synonyms if you have some score in a tag. But the synonym system is puzzlingly bad IMO; 4 people need to vote, but you don't see the suggestion unless you go on the tag page. See math.meta.stackexchange.com/questions/20956 $\endgroup$
    – glS Mod
    Sep 8, 2023 at 18:59
  • 1
    $\begingroup$ in practice, you can raise a flag on a question using the tag suggesting the synonym. If the need for it is clear we can manually create the synonym. If there plausibly might be controversy about it, I'd create a meta post asking whether others agree with the change (see eg this very post). To be clear, I'm totally fine with highly specific tags. We can have a tag for every qalgorithm under the sun as far as I'm concerned. But there needs to be a clear use case. If looking at a question you find yourself wondering whether a tag should be applied, then it's probably not a good tag imo $\endgroup$
    – glS Mod
    Sep 8, 2023 at 19:04
  • 1
    $\begingroup$ I agree about tags for every quantum algorithm under the sun. I have used them to search questions. It seems to me we should have fine grained tags for QIS and coarser tags for related fields (maths, computer science, physics, chemistry...). $\endgroup$ Sep 9, 2023 at 6:47

1 Answer 1

2
$\begingroup$

Following the discussion in the comments, a good compromise would be to have coarser tags for mathematical topics, such as linear-algebra, group-theory, functional-analysis, number-theory, analysis, etc.

There is the issue of what to do with the more generic that is currently used in a large number (currently ~500) of questions. However, given that removing it is a tall (and quite boring) order, we can just leave it, mark it as deprecated, and stop using it. Questions using it can be slowly retagged and so the tag can slowly go away (at some point, when fewer questions use it and we can thus ensure the safety of the operation, it can also be eliminated whatsoever eg merging it with another tag).

In practice, this change would entail removing most of the tags mentioned in the original question, except for linear-algebra (and possibly other overly specific tags that will appear or I missed here).

Please upvote this answer if you agree, and downvote if you don't (both are equally important). And of course, feel free to add here any comment relevant to the discussion or propose alternative solutions.

$\endgroup$
8
  • $\begingroup$ Thank you for writing this up! Few comments: I understand that the specific list of subfields of mathematics is not fixed and can be extended as needed. Examples we missed in the discussion so far: lie-theory and probability. I agree we should remove exponentiation and would add analysis or mathematical-analysis (potential synonyms). I think "calculus" just means "elementary mathematical analysis" so is probably not useful for us (since AFAICT we don't care about the distinction between elementary and advanced analysis). $\endgroup$ Sep 12, 2023 at 3:14
  • 1
    $\begingroup$ FWIW, I think that adding representation-theory could also be beneficial! $\endgroup$
    – Tristan Nemoz Mod
    Sep 12, 2023 at 11:36
  • 2
    $\begingroup$ @TristanNemoz I think that's also doable, yes. After all it appears often enough in QI and it's generally clear when a question focuses on it (eg all the questions about schur-weyl, many questions related to twirling channels or t-designs, etc) $\endgroup$
    – glS Mod
    Sep 12, 2023 at 12:21
  • $\begingroup$ @AdamZalcman yes analysis is certainly better. Though it probably wouldn't be used much, given most of the questions here end up being about linear algebra in some form or the other. Anyway, I wouldn't go about creating new tags here. This is more about deciding on a general line, and on which tags to remove. New tags (respecting the above guidelines) can then be created on a per-need basis as usual (I'll have to check whether removing specific tags might lead to posts requiring new ones; but I'd guess they'd just need linear-algebra in the vast majority of cases) $\endgroup$
    – glS Mod
    Sep 12, 2023 at 12:24
  • $\begingroup$ I agree we shouldn't create a tag unless we expect it to be used. It's quite possible that we don't actually need analysis. However, I think we should avoid the grab bag mathematics. I'm concerned that it's going to force us to clean things up later. It's also likely to be misused and overused. Instead, I'd either just remove exponentiation (and similar tags) without replacing them with anything or I'd replace exponentiation with analysis (and similarly for other narrow concepts). $\endgroup$ Sep 12, 2023 at 12:42
  • $\begingroup$ @AdamZalcman what do you mean with "avoid" mathematics here? As in, to remove it, or to leave it alone? It already exists and it's used in ~500 questions. I'd be fine with not having it, but removing it at this point wouldn't be easy $\endgroup$
    – glS Mod
    Sep 12, 2023 at 13:35
  • $\begingroup$ Removing or simply leaving as is (without increasing its use, e.g. as a replacement for exponentiation and the like) are both fine. We could also deprecate it and gradually remove it from questions that are edited for other reasons. $\endgroup$ Sep 12, 2023 at 14:03
  • 1
    $\begingroup$ @AdamZalcman ah, yes ok I see what you mean. That sounds good to me, I updated the answer to reflect this direction $\endgroup$
    – glS Mod
    Sep 12, 2023 at 14:13

You must log in to answer this question.

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