I know this question I'm proposing has some variants, but none had an answer that could solve the problem. Some others like this one: Here and Here

There are these questions (from newcomers who can never reach the search bar or the automatic results after typing the title) that appear from time to time. They are all the same because of some misunderstanding of very simple concepts and come in different forms.

I'm a Python guy and some examples are:

  • Mutable objects Vs. Immutable ones
  • Unbound Local Error (reference before assignment of objects)
  • Sorting/Ordering a dictionary (Hash tables shouldn't have a specific order)

There are so many variants of these questions and almost all of them have a lot of information missing, so that is hard to both give an appropriate solution for the person asking and to choose a good duplicate.

So, my proposal is:

Canonical answers to these questions. It could be done using Community Wikis but would be nice to have a special tag or symbol for them like [Canonical] on the title.

Doing something like that, the newcomers will get better answers and it'll be easy to find the "duplicate" question.


3 Answers 3


Perhaps, for users with very high rep in certain tags, they should be given the opportunity to act as the canon-makers. To their names would be applied a symbol similar to mods and chaos members. Perhaps the omega, to symbolize that they ordain a question as the last question ever needed for a particular subject?


They are given mod-like powers within a tag to immediately close questions as duplicates and to apply to questions which they deem to be the one true source of knowledge for a subset of similar questions. They would be allowed a dashboard that specifically targets questions flagged as duplicates for their tag, and to quickly review questions closed as duplicates by the common folk so that they may ensure the correct question was selected. From within this dashboard, they would be able to choose among all questions within their tag with for quick access when closing or re-closing a question

Another twist would be to allow users to apply a to questions, which the canon-makers would be able to review in their dashboard. They would either ordain a question as canon, or close them as dupes with similar questions.

This would have a number of benefits; we could offload lots of those "close as dupe" flags onto the Ω folks, use as a faq for a particular tag, etc etc. Linking to questions flagged as from within a tag wiki would also be nice.

The only problem I can see is how to deal with which tag the tag is associated with. If you have a question tagged

[c#] [.net] [derp] [canonical]

the canonical question could be about rather than . How this would flesh out I'm not quite sure...

  • 1
    Privileges tied to things other than rep levels have been talked about before, I think it was said to be a lot of work to implement. I'm not sure a tag would be the way to implement it, more like a status like closed / protected / locked, and a flag would be the way to suggest a question as canonical. That said, I like the general idea. I think 10k-20k rep + gold tag badge + some measure of community participation like two or three out of deputy / electorate / proofreader / archaeologist / copy editor would probably be the right criteria? It's quite a big feature request though.
    – agf
    Oct 6, 2011 at 17:39
  • 1
    I'm pretty sure that except for homework, tags that relate to the state of the question rather than content of the question have been avoided. Oct 6, 2011 at 17:52
  • 1
    Some users already have Ω in their display names
    – Flexo
    Oct 6, 2011 at 17:53
  • 2
    @awoodland: Then they will be shot. Pray, tell, what do you think stops users from adding a ♦ or a Ψ to their names?
    – user1228
    Oct 6, 2011 at 18:57
  • @ConradFrix: So, you're telling me that meta tags are discouraged, except when they are not? Spectacular observation, my dear fellow. Cheers.
    – user1228
    Oct 6, 2011 at 18:58
  • 1
    @Won'tಠ_ಠ - I'd assumed that ♦ would be filtered out of display names automatically
    – Flexo
    Oct 6, 2011 at 19:12
  • @Won'tಠ_ಠ Why yes, yes I am Oct 6, 2011 at 19:24
  • @Won'tಠ_ಠ I think that whatever your position on meta-tags in general or for FAQ questions, it works better as a status / flag thing than a tag thing :)
    – agf
    Oct 6, 2011 at 20:05
  • @agf: I'm not wed to the tag. We do have the whole post notice thing now....
    – user1228
    Oct 6, 2011 at 20:51
  • You might want to look at this one: meta.stackexchange.com/questions/68647/…
    – sbi
    Nov 20, 2011 at 0:32

The requires-no-extra-work-by-the-Stack-Exchange-team solution is:

  1. Find 10 or so of one of these questions. This should include already closed questions.
  2. Pick the one with the best existing answer. Preferably one already used as a dupe target.
  3. Rewrite the title of the question to be more general. Add searchable keywords / tags.
  4. Improve the answer if necessary.
  5. Vote to close the questions as dupes of the one you've improved
  6. Flag any already closed questions asking mods to re-open and re-close as a dupe of the new canonical question.
  7. Post the list here on Meta with a posse request. (Ping me somewhere.)

Once a bunch of questions have been closed as a dupe of the canonical one, it will come up as a close suggestion, and people will have an easier time finding it / remembering it.

I don't think new questions / community wikis / meta-tags are necessary.

For the does-require-extra-work-by-the-Stack-Exchange-team solution, see Won't ಠ_ಠ♦'s excellent suggestion (and my incisive comments).

  • Definitely the best idea. Oct 6, 2011 at 17:00
  • I don't think it's a job for a single person. That's why I prefer them as community wiki. The person doing that may not have all the knowledge needed for filling all the corner cases of some topic.
    – JBernardo
    Oct 6, 2011 at 17:04
  • 2
    @JBernardo the answer doesn't have to start out perfect. It will improve over time as people add to it, and naturally become a community wiki once ten people contribute. Just find a correct answer and work from there. It's a more natural process than what you suggest. I don't think the community creating new answers from scratch is going to happen (I'd be perfectly happy if the response to this question proved me wrong).
    – agf
    Oct 6, 2011 at 17:06
  • I like this answer a lot - it seems close to what I suggested at another Meta site: meta.gis.stackexchange.com/questions/3620/…
    – PolyGeo
    Aug 29, 2014 at 4:47

I think the problem in many of these cases is that people don't want to read the big canonical answer and figure out how to apply all the knowledge to their specific problem. The two areas that I see this mostly are in SQL and they are "How do I join these two (or more) tables" and "How Do Group By and Aggregate Functions Work". These are two questions that I see come up again and again. However, pointing them to a 500 word article explaining how joins work, or how group by works doesn't immediately solve their problem. Many people coming to this site want a very precise answer to a very specific problem, and aren't interested in educating themselves about the bigger picture. Sending a Newbie who knows nothing about SQL to a big article about joins won't help them solve the problem, and if confronted with such an answer, would probably find the site not helpful, and not return in the future.

  • 2
    Yes, I also think that if the person can't use the search bar (or google) he won't read a big article. I think more about a question with small answers (topics) about what's behind the problem.
    – JBernardo
    Oct 6, 2011 at 16:41
  • 1
    That's why there is a push for 3-4 canonical answers. For the Join topic, one would be: This is how to get the table together. Another would be: This is how to make sure the results don't appear more than once, because you didn't restrict the join predicate correctly. Another would be: This is how to change what coulmns are returned. Etc. Jun 15, 2012 at 12:30

You must log in to answer this question.

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