39

Now that we have one central cross-network Meta, we have a core place where network-wide issues can be discussed. There are a lot of great frequently asked questions and answers here, ones that will come up time and again on child Metas (like the new Meta.SO).

It would be great to be able to close these child Meta questions as duplicates of canonical entries here. This has come up a few times on Meta.SO already (such as for this question on Meta.SO, which is a duplicate of this question on Meta.SE), and we're ending up repeating ourselves by copying the answers from Meta.SE to answer the child Meta questions.

I know that there typically has not been an interest in allowing closing of cross-site duplicates as such, but this is a special case where I think there could be some benefit. It would both preserve one central repository of common canonical network-wide questions and answers and inform people of the presence of this site.

Otherwise, we'll be fragmenting the same questions across all of the child Metas, and in particular on Meta.SO we'll need to build up another set of canonical responses to the weekly "comments required for downvoting" and "why am I question banned" posts that we will certainly see so that we can close them properly.

2
  • 1
    I was just going to ask something similar to this after answering this question. Lol didn't realize you already linked to it
    – Travis J
    Apr 17, 2014 at 17:20
  • 2
    This isn't just useful for providing information. It is also important if these people have a unique view point; by adding that viewpoint to the the cannonical feature request they are more likely to have their view point in "the final mix" Apr 17, 2014 at 17:23

2 Answers 2

19

You can read my thoughts on cross-site duplicates in general here: The ability to link cross site duplicates

I understand the motivation for wanting to special-case metas, but honestly I think it ends up being worse.

Practical considerations

Let's imagine we did this, allowed entering meta.stackexchange.com URLs in the duplicate field on any question, added support for rendering these links in the title. We're still missing some major portions of what makes duplicates work:

  • Search doesn't find them. When you go to ask a question and the title-search fills in likely duplicates instantly, you're never gonna see the original - because it's on a different site. Same for the search in the "close as duplicate" UI, and, of course, the normal site-search.

  • The Related Questions sidebar doesn't list them.

  • The Linked Questions sidebar doesn't list them.

  • The Frequent Questions list doesn't list them.

  • ...you get the picture. There's a ton of infrastructure beyond just throwing a link at the top of a question, and none of it supports network-wide linking. It'd be handy if it did (in a lot of cases beyond just this) - but that's a much, much bigger change.

For the amount of support we could feasibly build into this, we're not really getting much in return beyond what we get by just posting an answer with a link in it. In fact, we're getting less than that, because with an answer we at least get the chance of offering site-specific guidance in addition to linking in the network-global discussion.

Philosophical considerations

Even if we managed to solve all of the practical problems, this still wouldn't be a great idea in many cases. Let's consider the single most-duplicated question on Meta: What can I do when getting "We are no longer accepting questions/answers from this account"?

This doesn't even apply to most sites. And the guidance in it is so specific to Stack Overflow in many areas that it's debatable whether referring folks elsewhere to it even makes sense. But without that guidance, it's not really useful to the folks on Stack Overflow, who make up the lion's share of all q-banned users. You see where I'm going with this? Even though it's a network-wide feature, the actual guidance for interacting with it must necessarily be customized to some extent per-site. There are a lot of questions like that...

My conclusion

MSE, like MSO before it, works best as source material: a comprehensive reference for how things work and why they work that way. It doesn't necessarily make for a good support tool without lots of hand-holding, which is tiring to perform every time there's yet another duplicate. Much better to have lots of partial-duplicates that reference it in the context of more localized problems than to close all those localized requests in hopes that each reader will slog through The Stack Exchange Bible in search of the information they need.

5
  • "we at least get the chance of offering site-specific guidance in addition to linking in the network-global discussion" -- this seals the deal for me.
    – Aditya
    Apr 17, 2014 at 17:50
  • 3
    Let's take the "people must be required to comment when they downvote" feature request as an example. The arguments are the same every time, as are the answers. There are some great explanations of this on the old question(s) on Meta.SE. Should we create a new canonical version of this question the next time someone asks on Meta.SO, copying across a lot of the responses that have been given? Would we then use that Meta.SO version as the one we point duplicates at? Apr 17, 2014 at 17:52
  • 1
    @BradLarson: On other communities (for example Ask Ubuntu), when we find that the question is a very basic FAQ question (how does bounties work, how does badges work, etc), we either only leave a comment on the question itself or provide an answer with to-the-point fact (yes, no, this, that) and provide the link to MSE for more elaboration.
    – Aditya
    Apr 17, 2014 at 18:03
  • 2
    I would summarize the responses in an answer, and link to the MSE version for backup @Brad. I've had to do this on many other meta sites - it's not that bad, and as I noted above it then solves the searchability problem for future readers.
    – Shog9
    Apr 17, 2014 at 19:02
  • There's one class of questions which could benefit, despite the drawbacks you just mentioned: feature requests. A fq might be languishing here, while gaining momentum elsewhere.
    – muru
    Nov 24, 2015 at 23:49
11

I fear that this ability would have a large potential for misuse. Many topics are intentionally allowed to be duplicated on per-site metas and MSE and should not be closed as cross-site duplicates in any case.

The only category where I don't see much value in duplicating are support questions about how the site works. Those have almost always the same answer across all sites (unless some site-specific configurations play a role).

Bugs should not be closed as cross-site duplicates as Shog and Tim Post already explained a few times today in various posts. The same also holds to a certain degree for feature requests. Discussions should also probably not be closed as the results can be drastically different depending on each site community.

My preferred solution would be to have a somehow curated list of FAQs on MSE that are automatically synched to each per-site meta in some fashion. There is really only a few dozen canonical questions that we need for this purpose. This would also mean that those questions are available via the search on each per-site meta, making it easier for users to avoid posting duplicates in the first place.

The exact implementation might be somewhat tricky, but I would prefer something like that to the rather problematic concept of cross-site duplicates.

You must log in to answer this question.

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