I searched something on Google. This is a screen shot of it.enter image description here

Link 01 : https://meta.stackexchange.com/questions/2974/is-there-a-way-to-remove-community-wiki-status/

Link 02 : https://meta.stackexchange.com/questions/2974/is-there-a-way-to-remove-community-wiki-status/83373

This is 100% duplicate articles. This will be a bad sign for Google.

But indexing this type of duplicate questions easily solved just adding following code for URLs like https://meta.stackexchange.com/questions/2974/is-there-a-way-to-remove-community-wiki-status/83373.

<link rel="canonical" href="https://meta.stackexchange.com/questions/2974/is-there-a-way-to-remove-community-wiki-status" />
  • 1
    I think this is obvious issue. Either there should be 1 result, or the answer's link should have a prefix [Answer] or like that, as user can differentiate... BTW, i think the link to question is enough, no need separate link to answer, so answer's link should have canonical meta tag i think, refering to main. – T.Todua Apr 2 '18 at 11:20

The way question URLs work is that anything after questions/{question-ID}/ is dynamically replaced with the title of the question... But it looks like any string added after that title (following another forward slash) in the URL is kept in the URL itself (i.e. it isn't dynamically removed as it should be) and added to the canonical link too.

Compare the question with only the title:

enter image description here

...with the extra string:

enter image description here

...and another random question with a completely random string:

enter image description here

There should only ever be a single canonical URL for a question so this is certainly a bug. I assume the fact that it isn't removed from the URL in the first place is the bug, but if there is a reason for that then the canonical link most certainly is.

What I assume happened is that someone posted a link to that question somewhere with the extra bit on the end of the URL, Google then crawled that and saw it as canonical so indexed it along with the normal URL.

  • That second link is a link to the answer. This may be intended. – ale Jun 26 '17 at 13:25
  • Hmm, yeh, without the the URL fragment (#1234) though it just links to the question... and since they're all on the same page there should only be a single canonical link – Cai Jun 26 '17 at 13:29
  • 1
    Heh I was just gonna comment above I'm sure we add canonical URLs, but apparently it's broken... The last ID number represents the ID of the answer you're accessing on the question, normally follow by #{same-answer-id} which is used to scroll to and highlight it in JavaScript. I don't remember why the one that's part of the path exists, but it's supposed to be there. Probably for tracking since the server doesn't get the fragment. – animuson Jun 26 '17 at 14:32
  • @animuson yeh I don't see any obvious reason for the ID to be added to the path (it does nothing without the # fragment), but it shouldn't be added to the canonical (and there should probably be a check to make sure the ID actually corresponds to an answer) (oh, I didn't see your edit—tracking makes sense) – Cai Jun 26 '17 at 15:09
  • Right now there is no canonical tag at all anymore on all Stackexchange sites. Why? – user1666456 Feb 3 at 10:41

It is an issue with google.

As the title of the question edited, the link will change.

But FYI, even https://meta.stackexchange.com/questions/2974/abcde will redirect you to the question. So, google think the link is still live(It is true. Even it is changed, the old link is live), it won't remove it from it's directory and add the new one.

So, https://meta.stackexchange.com/questions/2974/abcde and https://meta.stackexchange.com/questions/2974/abcde-efgh both will be there.

The removal is somewhat difficult part. Either provide a 301 redirect or disallow in robots.txt and such methods are there. But for such a heavy site, with thousands of title edits per day, it is difficult to bring it in action.

  • What you tell is a different thing. This is not about changing title and then automatically change the URL. What I suggest should not be hard because current link will be in database anyway. So canonical links should be to that link. – Damith Ruwan Jun 27 '17 at 6:00

You must log in to answer this question.

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