3
$\begingroup$

I recently proposed (Do we need tags for highly specific elementary mathematical topics?) we cut down on the number of tags on highly specific mathematical topics.

However, I've thought for some time that there's a significant number of issues with the tags we have. Rather than creating a post for each one individually, I figured it might be better to have a single post to collect these, and more handly collect opinions about each proposal.

I'll start by adding the ones I'm aware of, and that I most often found frustrating to deal with when deciding which tags should be used in a certain question (and as you can see, there's already quite a few of these). Please feel free to add other answers about any other issue (related to tags) that you think needs fixing. And remember to upvote/downvote on each proposal you agree/disagree on, and add comments to voice disagreement or suggestions to modify a proposal, or anything else.

I'm featuring this post to give it a bit more exposure, in the hope of enough people seeing it to create a meaningful volume of discussion, and maybe also capturing some other issues other people might have seen and tag-related proposals they might have.

After this is done, I also think it will be a good idea to have a general discussion on tag management, to avoid having to do this again in the future. I'm thinking another meta post where we decide on general guidelines of which tags we should and shouldn't allow. More specifically, I think it's a good idea to have guidelines of the form "users should be allowed to create tags of type X and Y as usual, whereas tags of type Z and W should require a tailored meta discussion beforehand", where X and Y might include for example "tags about specific quantum algorithms" and Z and W might include for example "tags about specific classical programming languages". Anyway, that's for later.

EDIT 14/10/2023 — I think this has been up for a sufficiently long time, so I'll start rolling out the upvoted proposed changes.

$\endgroup$
1
  • $\begingroup$ "I'll start rolling out the upvoted proposed changes" -- please don't be ridiculous. Just because something has an upvote, doesn't mean that you have the community's permission to do anything. Upvotes can come from anywhere, and many people (including me) do not like to downvote, so even though I don't like some of your proposals I have not downvoted them. Please make decisions based on actual consensus rather than just blindly looking at upvotes (without even any consideration for how few upvotes there are!). $\endgroup$ Oct 14, 2023 at 19:02

10 Answers 10

6
$\begingroup$

Clean up tags related to ibm. We currently have

Do we really need all of them? Here I'm also asking because someone else might be more familiar with this than I am, but isn't it sufficient to have a single tag (eg or maybe ) for all of these?


Implemented: (1) and now point to ; (2) points to .

$\endgroup$
6
  • $\begingroup$ I hope we'll have an input from an IBMer on this one, but AFAIK ibm-q-experience and ibm-quantum-experience are perfect synonyms. The latter having only 9 questions, I think it's safe to outright remove it and retag these questions. The ibm tag guidance explicitely says that it's about their devices and should thus point to ibm-quantum-systems. However, the former is way more (mis?)used, and the latter has no tag guidance. I think it should be in this direction since ibm isn't really verbose. I do think it should stay separate as it targets IBM's devices, not their platform. $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 8:58
  • $\begingroup$ Finally, it seems like ibm-quantum-services is already pointing to ibm-q-experience if I'm not mistaken? $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 9:01
  • $\begingroup$ Trying to fit too many info on a comment wasn't the best idea I've had, but basically: 1. ibm-quantum-experience -> delete and retag 9 questions with ibm-q-experience 2. ibm -> ibm-quantum-systems, since they have the same tag guidance and the latter is more verbose. Also, it should stay separated from the ibm-q-experience tag since they target quite different topics 3. It seems like ibm-quantum-services -> ibm-q-experience is already in place, which I think is what should be $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 9:05
  • $\begingroup$ my bad about ibm-quantum-services, I didn't notice it was already a synonym. So in summary, you're saying ibm$\to$ibm-quantum-systems and ibm-q-experience$\to$ibm-quantum-experience? That's fine, but for constistency I'd spell out the quantum in both (or alternatively use q in both). We can just convert all ibm-q-experience to ibm-quantum-experience if we want, that's easy. Another thing: if ibm-quantum-systems is about the devices, might it be better to rename it to ibm-quantum-devices though? $\endgroup$
    – glS Mod
    Sep 18, 2023 at 10:38
  • $\begingroup$ Concerning the "Quantum Experience", when it was first introducted, it was named IBM Q Experience, so I don't think we should change the Q into Quantum since this is the name people are used to (though I struggle to find any reference to this name on the IBM website today, so I'm not even sure it is still used tbh). Thus, I would be more in favor of deleting ibm-quantum-experience and retag the 9 questions with this tag to ibm-q-experience. I'd vote for changing systems for devices indeed! $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 11:24
  • $\begingroup$ I suggest to keep ibm and ibm-quantum-experience separate. I suggest to point ibm-q-experience to ibm-quantum-experience. As for ibm-quantum-systems, I do think a separate tag for the various different chips is appropriate, but this doesn't have to be unique to IBM. For example, questions about Google's Sycamore or Bristlecone architectures, and questions about the IBM Tokyo and IBM Melbourne architectures, can all go under a quantum-chips or quantum-processors or specific-quantum-architectures tag, or something similar. $\endgroup$ Oct 7, 2023 at 1:02
5
$\begingroup$

Synonymize and . I feel like the difference in meaning between these is small enough that having them both creates more confusion than anything else.

I'd lean towards having point to , but I'd be fine with the other way around. We could even just have both point to .


Implemented: both and now point to .

$\endgroup$
2
  • 4
    $\begingroup$ I like terminology-and-notation since it explains the intention here better. $\endgroup$ Sep 16, 2023 at 23:19
  • $\begingroup$ Terminology and notation aren't exactly the same. What would be the purpose of combining them? $\endgroup$ Oct 6, 2023 at 11:47
5
$\begingroup$

Remove highly specific mathematical tags (this is essentially Do we need tags for highly specific elementary mathematical topics?, I'm just repeating it here to have all tag-related proposals in one place).

More specifically, remove , , , , , , .

Keep , and if needed allow other more coarse-grained mathematical tags such as linear-algebra, group-theory, functional-analysis, number-theory, analysis, representation-theory, etc.


Implemented: manually removed from the few questions that used it; [inner-product], , , , now point to . Left alone, at least for now.

$\endgroup$
1
4
$\begingroup$

We currently have , , .

I'd at least have point to . But even better, I think it would cause less confusion to have both point to , maybe changing its name to .

Implemented: all the above are now synonyms of .

$\endgroup$
3
  • $\begingroup$ +1 except I'd remove applications from this list. In my experience, most folks use the term "applications" to refer to computational tasks that present or future quantum computer could do, e.g. quantum simulation or topological data analysis, and not to experimental and engineering work related to the attempts to build a quantum computer. Current use of the tag mostly reflects this. $\endgroup$ Sep 16, 2023 at 23:15
  • $\begingroup$ @AdamZalcman that's fair. Edited. Thanks $\endgroup$
    – glS Mod
    Sep 17, 2023 at 13:59
  • $\begingroup$ Not all experiments are about physical qubits. I think these ones can stay as they are. $\endgroup$ Oct 6, 2023 at 11:38
3
$\begingroup$

I'm not sure how and are relevant, when already exists. None of them have tag guidance and the latter doesn't have any question associated with it.

I do get the point of these tags (especially the latter, I do think primitives are an important topic in Qiskit and understand how it could help beginners), but I fear that leaving them is opening some kind of Pandora's box.

Since they target kind of specific Qiskit features, should we do the same for all the features of all the other frameworks? I could hear that the balance of questions between Qiskit and the other frameworks isn't the same, but it's not like these two specific tags are widely used anyway, which could justify a special place.

Thus, I would be in favor of deleting both these tags.


Implemented: and now both point to .

$\endgroup$
3
  • $\begingroup$ ah, I knew I had forgotten something, thanks. I agree with this. I don't know much about qiskit-primitives to tell whether it's worth having it though, so I'll just trust the judgments of you guys that mostly interact with qiskit-related questions. On a related note, what about qiskit-runtime? Is that worth having? $\endgroup$
    – glS Mod
    Sep 18, 2023 at 10:34
  • $\begingroup$ @glS Ah, that's a tricky one. On the one hand and contrarily to qiskit-primitives or qiskit-aer, we can (I think) consider it as separate thing from Qiskit. It basically relates to running Quantum Circuits on IBM's Quantum machines AFAIK. However, it suffers from another problem in my opinion, which is that it's not quantum-related, and we still have this weird dichotomy between SO and here for programming-related questions. $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 11:32
  • $\begingroup$ The different situations I can imagine for a question using this tag are 1. The code works offline but not online 2. The fact that it's using runtime is not relevant (this seems to be the case about the first few questions using this tag I've looked at, which isn't a significant sample). In both cases, I'm not sure we shoup keep this tag (better answered on SO or not useful for answering the question), but it's already a bit used, so I don't really what to do. Don't hesitate to ask for details on runtime if you want more info, I'm definitely not an expert on the topic but I'll try to help! $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 11:36
3
$\begingroup$

I inadvertently and imprudently created . I think this should be merged with .


Implemented: the tag has been removed and marked as deprecated in the info.

$\endgroup$
3
$\begingroup$

Remove tags about specific (classical) programming languages. I'm essentially referring to (unless there's others I haven't seen).

I don't think we should have these. We have to encompass the general brand of questions focusing on the programming aspect, but I think having things like encourages people to ask questions about a (classical) programming language itself, and such questions should be asked on StackOverflow, not here.


Implemented: now points to

$\endgroup$
2
  • $\begingroup$ I think it would be better to keep the python tags for questions that are about python. $\endgroup$ Oct 6, 2023 at 11:37
  • $\begingroup$ I think that going ahead and implementing this without reaching community consensus was a bad idea. Python is a great tag, and I believe that it belongs on this site. $\endgroup$ Oct 14, 2023 at 18:51
3
$\begingroup$

Invert the order of and . Currently points to , and I'm proposing to have it the other way around.

This is a minor one and mostly "aesthetic", but I think is more explicitly about non-classical algorithms. We might use a separate tag for classical algorithms if needed be.


Implemented: renamed into , and removed the synonym. We'll thus have to manually take care of retagging eventual questions that will use .

$\endgroup$
10
  • $\begingroup$ I think algorithm is potentially confusing and would support its removal. Another interesting related tag here would perhaps be syndrome-decoding-algorithm for questions about classical algorithms used to diagnose errors in quantum error correcting codes (e.g. MWPM-based algorithms for the surface code). $\endgroup$ Sep 16, 2023 at 23:24
  • $\begingroup$ to clarify, you mean to leave only quantum-algorithm and remove algorithm altogether (thus without making it a synonym for quantum-algorithm)? $\endgroup$
    – glS Mod
    Sep 17, 2023 at 13:57
  • $\begingroup$ Yes, I think algorithm is too generic to be useful. $\endgroup$ Sep 17, 2023 at 13:59
  • 1
    $\begingroup$ @AdamZalcman the thing is: if you synonymize them, if you start typing algorithm you'll get the prompt to use the tag quantum-algorithm (precisely the opposite of what happens now if you try using the tag quantum-algorithm). So effectively this would mean removing algorithm, and at the same time preventing someone to recreate it. If you instead only merge the tags (essentially retagging algorithm$\to$quantum-algorithm) that doesn't prevent people from still using algorithm in the future. $\endgroup$
    – glS Mod
    Sep 17, 2023 at 19:00
  • $\begingroup$ I worry that the distinction that the algorithm tag hides, especially quantum-vs-classical you mentioned in this answer, are actually very relevant to the users. As a result, any substantial use of algorithm will probably lead to a future need for another cleanup wherein we'll find ourselves replacing it with more specific tags, such as quantum-algorithm, classical-algorithm, shor-algorithm, syndrome-decoding-algorithm etc. $\endgroup$ Sep 18, 2023 at 7:10
  • $\begingroup$ @AdamZalcman what about adding a classical-algorithm tag then, and having algorithm point to that one? I just think it would be better to not leave a possibility for a user to use algorithm to mean quantum-algorithm. Synonymizing it to classical-algorithm should make its meaning much clearer. This might be useful for things like discussing classical algorithms relate to quantum ones. Eg things like reducing factoring to period finding $\endgroup$
    – glS Mod
    Sep 18, 2023 at 10:42
  • $\begingroup$ I view algorithm a bit like mathematics. Both are too general to be useful for searching and are likely going to be misused and require future cleanup. For these reasons, I'd prefer them being removed rather than synonymized to something else. However, if algorithm is to remain, then I think your original proposal is better than synonymizing to classical-algorithm since I expect the letter to be less frequently used than quantum-algorithm. $\endgroup$ Sep 20, 2023 at 1:27
  • $\begingroup$ @AdamZalcman I think that's fine. So in conclusion, you're saying to merge algorithm$\to$quantum-algorithm, and then just leave algorithm be. But just to clarify: a tag "existing" is a bit of a vague concept here. The tag won't be used by questions after the merge, but that doesn't prevent future users from adding it to new questions. That would require CMs intervention, us mods here can't do it directly. Still, we can just merge it and then manually enforce its non-usage $\endgroup$
    – glS Mod
    Sep 24, 2023 at 0:34
  • $\begingroup$ Ah, didn't realize this. Thanks for the clarification! $\endgroup$ Sep 24, 2023 at 0:37
  • $\begingroup$ If we do this, then many non-quantum algorithm questions will be tagged as "quantum" algorithm. $\endgroup$ Oct 6, 2023 at 11:41
-1
$\begingroup$

This question discussed with - to me, they are different enough topics but I agree that is less niche with potentially broader appeal.

$\endgroup$
5
  • $\begingroup$ When did I suggest that? Also, the @ping function works in comments and chat, not in questions and answers, so I didn't get a notification about this ping :) $\endgroup$ Oct 6, 2023 at 11:45
  • $\begingroup$ @user1271772 sorry for not pinging you! I read your answer here as suggesting that! $\endgroup$ Oct 6, 2023 at 23:41
  • $\begingroup$ Okay, it makes a lot more sense now that I see it in the context of that question! Yea quantum-money still has only 9 questions (but 4 watchers!), and cryptocurrency only has 13 questions (and 1 watcher). Cryptography on the other hand has 123 questions and 12 watchers. I don't see a huge problem with keeping all three tags, but also, tags with only a dozen or so questions in the span of 5 years will probably never get to the stage where they will help much for moderation purposes (for example people that earn the gold badge on a particular tag, can hammer close questions with that tag). $\endgroup$ Oct 7, 2023 at 0:53
  • $\begingroup$ Yeah certainly! I was just excited five years ago because I finally understood the Quantum Money from Knots paper well enough to ask a question that John Watrous of all people answered so I thought I'd be hot stuff to propose the tag. But I totally agree that it's a pretty hyper-niche tag either way. $\endgroup$ Oct 7, 2023 at 1:03
  • $\begingroup$ It might be a hyper-niche tag, but I don't see a problem with keeping it. It's a sub-field of its own, and there's 4 people "watching" the tag. One of the things that I think is most important when considering whether or not to delete a tag, is the number of people "watching" the tag. Those people could also be subscribed to get emails whenever a new question is asked with that tag. If John Watrous is "watching" the quantum-money tag, we don't want him to suddenly stop getting emails one day, just because we suddenly deleted the tag. $\endgroup$ Oct 7, 2023 at 1:09
-2
$\begingroup$

Synonymize and (keeping the former).

I don't see the need to have both. They are mostly "filler tags" in my opinion, but I can't think of an example where it's useful to have density-matrix as a tag. If the focus is a quantum state, we can just use . Whether the quantum state is represented as a density matrix or not is hardly of substance.

An exception might be questions about elementary topics that are asking specifically about properties of density matrices. But even then, this seems such a narrow topic that we might as well tag these questions with and .

$\endgroup$
9
  • 1
    $\begingroup$ Can you run any analytics between the two, referenced against the reputation of the OP? I would assume that quantum-state is used as a placeholder for new users while density-matrix may be used by questioners with higher rep. $\endgroup$ Sep 17, 2023 at 21:59
  • $\begingroup$ I may be biased as someone having asked basic density matrices questions, but I think I'd like to keep them separate. As you've said, the questions specifically about density matrices would in general be elementary questions, but I do think it helps having a separate tag so that people can search within it if they want to learn about this topic. Also, I don't think the same reasoning as for tags like inner-product applies here, as AFAIK density matrices are a QIS-specific tool (i.e. this would not be questions better asked on Math.SE) $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 8:43
  • $\begingroup$ @TristanNemoz I agree the reasoning is different than the one for basic mathematical tags. But even if the goal is having a tag to categorize elementary questions about density matrices, I'm not convinced density-matrix actually succeeds in doing that in practice. Just have a look at the questions that end up using it. There's currently 336 questions that use it; how many of these are actually about properties of density matrices per se, and how many are using it because the word "density matrix" just happens to appear at some point in the question? $\endgroup$
    – glS Mod
    Sep 18, 2023 at 8:49
  • $\begingroup$ I really think the ease of maintenance always needs to be taken into serious consideration for tags. If a tag leaves room for being misinterpreted/misued by a user that is not familiar with the site's rules, that means we are going to have to enforce the correct usage by manually retagging questions. And if we are talking about a tag as used as density-matrix, that means potentially a lot of manual retagging (including retagging many of the ~300+ questions that currently exist). Is it worth it just to have a nice list about basic properties of density matrices? $\endgroup$
    – glS Mod
    Sep 18, 2023 at 8:53
  • $\begingroup$ Well, I think that the interests of having this tag are twofold: 1. It eases the search for elementary questions about density matricse for newcomers 2. It adds some context to the question. If I see that density-matrix is specifically added along quantum-state, I know the type of answers I should give and I know that the question won't probably deal with pure states (which generally is also a huge hint about the question's "sub-field", like most of the algorithms I know of deal with pure states). $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 9:13
  • $\begingroup$ Of course, I'll follow the community's opinion! But my take on this is that they should stay separated. $\endgroup$
    – Tristan Nemoz Mod
    Sep 18, 2023 at 9:14
  • $\begingroup$ @MarkSpinelli apparently I can (thanks GPT4!). Getting the list of reps from data.stackexchange.com/quantumcomputing/query/1785197/…, and some quick post-processing with MMA, I get the following (renormalized) histograms for the reps of askers for the two tags: i.stack.imgur.com/B5jwW.png. I'm truncating reps $\ge5000$ as there's only few of those and it hides features about relatively lower rep users. Axes are probability of bin (ie fraction of questions that have been asked by users with rep in the bin) vs rep. $\endgroup$
    – glS Mod
    Sep 18, 2023 at 10:12
  • $\begingroup$ Not sure what we can conclude from this though. We can see that quantum-state is comparatively more used by lower-rep users, and we can see small bumps attributable to individual high-rep users (eg I think the one at ~5000 is just @TristanNemoz, and there were similar ones attributable to me and others at higher reps). But my main doubt remains how to clearly distinguish the two, and how many of the questions currently using density-matrix should just be using quantum-state instead $\endgroup$
    – glS Mod
    Sep 18, 2023 at 10:13
  • $\begingroup$ Quantum state could be about wavefunctions or Wigner distributions, I think density-matrix should be kept. $\endgroup$ Oct 6, 2023 at 11:43

You must log in to answer this question.

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