2

2 suggestions for canonical questions:

  1. The recently established canonical question for field switching after PhD has the following added to note to close voters:

    Questions about switching from X to Y may be closed as a duplicate of this. However, please be sure that the below answer actually answers OP's question before voting to close! If the question asks about something not covered below (other than the specific fields), the question should be left open until its answers are merged into this canonical question.

    I think that is a great addition and should be included in all canonical questions, as I have the feeling that often, questions that kind of fall into canonical territory get closed even though the specific answer is not (yet satisfactory) included in the canonical answer (the journal workflow is one of those).

  2. Although canonical questions do exist, they are not easy to find as the float around in the sea of questions indistinguishable from other, non-canonical questions unless you actively click on them. My suggestion for making the canonical questions more "findable" would be to simply create a tag named canonical and ad this tag to all the canonical questions, with an explanation in the tag wiki on how canonical questions work, how people are invited to add to them etc.

4

5 Answers 5

3

Thanks for your suggestions. I'll write my responses in two different answers, so people can upvote/downvote individually.

I agree we should add some version of the quoted text to each canonical question. Naturally, we'll have to think a little bit carefully about phrasing these: we should try to be clear about which questions are merely "specific instances" of the canonical question (and should be closed) and which are asking for general information that should be in the canonical answer but isn't.

Along similar lines, I like how the newest canonical question's "notice to readers" paragraph gives a brief explanation for why questions about switching from X to Y are now considered duplicates of the canonical question; this should reduce confusion / annoyance when someone's specific question is closed as a "duplicate" of the much broader question. We should make sure the other canonical questions have something similar.

3

We already have a list of canonical questions. The advantage of this (over a tag) is that it can be sorted by category and similar. The obvious problem is that users are not sufficiently aware of this list or otherwise don’t know where to look for canonical questions.

If we add a banner to canonical questions, we can also include a link to this list. This should obviate the need for a specific tag, as it is then visible on every canonical question, just like the tag. This way we avoid the problems of a tag, like mistagged questions, the tag being removed, etc.

8
  • This is certainly better aligned with the "StackExchange Philosophy." On the other hand, the complaint seems to be that canonical questions are hard to find; adding a hard-to-find link to the hard-to-find pages may not solve the problem. (Personally, I am agnostic; having bookmarked the canonical list in my browser, I never really experience the purported issue).
    – cag51 Mod
    Commented Aug 3, 2022 at 18:42
  • But how would a tag solve this problem better than having a link in the canonical-questions banner? In both cases, you need one canonical question as a gateway (to the tag or meta list). But then maybe I am bad at putting myself into the shoes of those who have problems finding the list.
    – Wrzlprmft Mod
    Commented Aug 3, 2022 at 21:05
  • 1
    With a tag, you could click on the "tags" button on the left and then navigate to the canonical-questions tag...which is maybe a bit less hacky than trying to locate either the list or an existing canonical question, which for me usually involves using google and site:academia.stackexchange.com. The mobile experience may play into this too, not sure (I rarely do more than read the site on my phone).
    – cag51 Mod
    Commented Aug 3, 2022 at 23:14
  • IMO, the banner is better than the tag for at least one reason, we only have 5 tags per question. For example, buffy edited the workflow of journal question by replacing the peer-review tag with canonical question. However, peer-review tag is very important for this question. Do we really want to sacrifice peer-review for canonical-question tag?
    – Nobody
    Commented Aug 4, 2022 at 13:04
  • Another thing, I always use search to find that journal workflow question by searching for "workflow". There are not many "workflow" on our site, the search is relatively quick. Just FYI.
    – Nobody
    Commented Aug 4, 2022 at 13:07
  • @Nobody, I thought pretty long about which tag to drop and decided that peer-review would most likely have been combined with other of the remaining tags, so would be minimally disruptive.
    – Buffy
    Commented Aug 4, 2022 at 15:08
  • 1
    The problem with banners is that you first need to find the question in order to see the banner. Those who retag questions considered canonical already know about canonical questions and so don't really need the banner.
    – Buffy
    Commented Aug 4, 2022 at 15:11
  • @Buffy I am thinking for the first time users rather than >3k voters.. If someone sent a manuscript to a journal and keep getting "with editor" status, they come to us to find information. It is possible they think it's an faq. It is also possible they think it's about peer review. I don't think they would think it's about a "canonical question" because it's not very often used on other sites. So, I think if we really want to use a tag, "faq" is better than "canonical-question".
    – Nobody
    Commented Aug 5, 2022 at 4:55
0

Thanks for your suggestions. I'll write my responses in two different answers, so people can upvote/downvote individually.

I agree the "canonical question" tag is probably a good idea. But, there are some potential issues we should be aware of.

First, meta-tags have been banned from Stack Overflow and are explicitly discouraged elsewhere. Looking through the rationale, many of the concerns would not apply to a tag, but this ones does: "The reason meta-tags are a problem is that they do not describe the content of the question. They describe some other aspect of the question, like the author’s skill level, or the author’s motivation for asking it, or generally what 'kind' of question it is (poll, how-to, etc.)." I think creating a singular meta-tag is probably not a terrible idea notwithstanding the above, but we should be clear that this is a singular event and not a precedent.

Second, there is concern that novice users will misuse this tag and make random questions "canonical." Mod-only tags were suggested a year ago; I think mod-only tags are a fine idea, but they would have to be implemented by the StackExchange developers, which would take years (if they agree to do it at all, which I think is unlikely given the above). So, we would have to manually detect and remove incorrect applications of this tag; this is a bit of a maintenance burden (though I think it's probably worth it in this case; people have been complaining about how difficult it is to find canonical questions for years).

6
  • 1
    Note that I review some tags daily, especially New tags and do a fair amount of cleanup. I'm happy to continue this for the canonical-question tag and remove it when I think it is being misused (as I do for other tags BTW). We can seek a long term solution (and should), but a bit of diligence by a few users in addition to mods can help keep the site clean. It is probably a mistake to add canonical-question to more than a few questions per day to avoid flooding the main question page. And a review of my wiki entry for canonical-question should be done by interested parties.
    – Buffy
    Commented Aug 2, 2022 at 12:44
  • I would also volunteer to keep an eye on a tag like this, it would probably suffice to check once a week or so.
    – Sursula
    Commented Aug 2, 2022 at 14:37
  • 1
    @Sursula-they-, we get a lot of questions in a week. You can add the tag to your "watched tags" list and it will become obvious on your screen when a question appears with the tag.
    – Buffy
    Commented Aug 2, 2022 at 14:50
  • Note that while it’s easy to track the tag being added to a question by subscribing to the tag, that does not apply to the tag being removed. CC@Sursula
    – Wrzlprmft Mod
    Commented Aug 3, 2022 at 17:43
  • I think, with some experience, that implementing mod-only tags should not take years. Most of the code is already in place for other purposes. Testing would be required, of course. But "it would take years" reflects very poorly on the skills of the developers. Have the mods here requested it as a feature? If so, what was the result?
    – Buffy
    Commented Aug 4, 2022 at 17:54
  • My time estimate is "data-driven" -- the feature announcements coming out on main meta now were requested 5+ years ago. Once they actually start working on it, I agree it should be very easy/fast to implement, but there seems to be a long and slow-moving queue. Not sure if anyone on our team has requested it -- I personally have not, because I think they will altogether refuse, citing the above blog post. But I'd love to be proven wrong. Also, AFAIK, there is no special way for mods to request features, it's just a matter of creating a post on main meta.
    – cag51 Mod
    Commented Aug 4, 2022 at 18:44
0

canonical questions do exist, they are not easy to find

My suggestion is: This is not a bug. Canonical questions are the same as every other question, except that they have been discussed on the meta site. They do not need special treatment.

The good ones are easy to find because there are many duplicates pointing to them.

-2

However, please be sure that the below answer actually answers OP's question before voting to close!

This is wrong.

  • When voting to close, you are voting for a duplicate question, not a duplicate answer.
  • If two questions are the same, or should have the same answer, then they are duplicates, even if no answers exist in the site. This rule should apply to all questions, canonical or not.
11
  • This doesn’t work well for canonical questions. Consider, for example, What does the typical workflow of a journal look like? How should I interpret a particular submission status?. This Q&A gives a (hopefully good) overview that answers many basic questions (which is why we created it), but it cannot possibly answer every question within its scope. If we close questions that could theoretically be answered by this but aren’t, those will likely never be answered.
    – Wrzlprmft Mod
    Commented Aug 13, 2022 at 21:11
  • Also, SE duplicates are more answer-focused than this answer presents it. If an existing Q&A completely covers another question, the latter is a duplicate, even if the question is technically different. In fact, this applies to many questions we rightfully close as duplicates of canonicals. On the other hand, there is a reason that you can only flag/vote to close a question as a duplicate if the target question has an answer.
    – Wrzlprmft Mod
    Commented Aug 13, 2022 at 21:16
  • "If we close questions that could theoretically be answered by this but aren’t, those will likely never be answered." That is exactly what has been happening. The solution is to read the questions being closed and edit the answer to the canonical question. Commented Aug 13, 2022 at 22:30
  • "If an existing Q&A completely covers another question, the latter is a duplicate, even if the question is technically different." Could you explain how this can happen, if the answers are all on topic? Commented Aug 13, 2022 at 22:32
  • "there is a reason that you can only flag/vote to close a question as a duplicate if the target question has an answer." Are you sure? I thought I had done that before. Commented Aug 13, 2022 at 22:33
  • That is exactly what has been happening. – Yes, and it has not always been good. — The solution is to read the questions being closed and edit the answer to the canonical question. – This is occasionally sensible, e.g., if you have a submission status that is not listed in the canonical yet but an obvious analogue of a listed status. But quite pragmatically it doesn’t work as you idealise it: Often we have questions closed as duplicate where the answer would be too special for the canonical or nobody edits the canonical.
    – Wrzlprmft Mod
    Commented Aug 14, 2022 at 6:17
  • Here is an example for the above.
    – Wrzlprmft Mod
    Commented Aug 14, 2022 at 6:40
  • Could you explain how this can happen, if the answers are all on topic? – For a simple example, consider: Should I get a letter of recommendation from my mother, who is a famous researcher in my field? and May I write a letter to support my girlfriend's application? The relationship between recommender and recommendee is different; one is from the recommender’s and one from the recommendee’s perspective. Yet what answers one question also answers the other; they are duplicates.
    – Wrzlprmft Mod
    Commented Aug 14, 2022 at 6:51
  • "there is a reason that you can only flag/vote to close a question as a duplicate if the target question has an answer." Are you sure? I thought I had done that before. – See this FAQ.
    – Wrzlprmft Mod
    Commented Aug 14, 2022 at 6:56
  • @Wrzlprmft I do consider those two questions to be duplicates, even if they are technically different, but my viewpoint is not based on the existing answers. Commented Aug 14, 2022 at 12:22
  • The example did clarify your viewpoint. Commented Aug 14, 2022 at 12:23

You must log in to answer this question.

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