A partial archive of meta.discourse.org as of Tuesday July 18, 2017.

Webmaster is giving errors of Duplicate Meta Descriptions and Duplicate Title Tags


Some of our forum users are not able to write a precise and specific title for Topics they publish, therefore most often we re-edit the titles of the topics to give a better understanding of the thread discussion.

Discourse helps re-editing the title completely without effecting the URL location. It simply changes the permalink without effect the URL location.

Both the following URLS will point you towards a common page




But the downside is that Google webmaster tool is considering both the URLS as two separate pages and considering it as a mistake. I have now around 114 Duplicate descriptions

and 16 Duplicate Title tags because of the parameter ?page=2

I am looking for a quick fix for these problems and how can we make sure google stops considering them as two separate pages when in fact they are pointing to the same location? confused


Hmm, is there a <link rel=canonical included on the pages?


Of course all pages has it but who would tell that to Google?


IIRC, Google invented it.

The problem is each ?page gets it's own canonical but all have the same title eg.

<link href="http://community.sitepoint.com/t/the-person-above-you/1836?page=1252" rel="canonical" />

Hey, ?page= links shouldn't be being used there! Canonical links are of the /t/slug/topicid/postid form.


?page= should only be used if you are also using rel=prev and rel=next.

Keep in mind that Google gets a static HTML version of discourse, which does not have infinitescroll loading. In fact it doesn't have any pagination logic on it at all. So really to Google the first page and any other pages for the same thread would conflict with each other or be considered separate pages entirely.

For the no Javascript rendering there should be a next/previous button for pagination at the least, and in the head, a rel pref/next meta tags for tracking pagination. Google then would be able to index long threads and attribute all of the content to that one thread. More info: http://googlewebmastercentral.blogspot.com/2011/09/pagination-with-relnext-and-relprev.html

This would also solve the duplicate title issue so long as the slug was the same in the canonical.


I agree with you @chapel I guess you provided the most accurate solution to this ongoing problem. I hope @codinghorror considers it seriously because it is a serious SEO mistake in the noscript version


sure happy to amend this just outline exactly what needs doing, with sample HTML


@sam Also note that the About page and Homepage are have same description thus again giving a Duplicate meta description error.


If you want, next week I can do a PR to implement some of the SEO best practices. As far as what I can do without touching the Ruby.

Putting it on my calendar.


Here's what I see when I use Firefox with my user-agent set to googlebot:

<div role='navigation' itemscope itemtype='http://schema.org/SiteNavigationElement'>
    <span itemprop='url'><a href="/t/show-desktop-notifications-when-available/10183?page=2" itemprop="name" rel="prev">? previous page</a></span>
    <span itemprop='url'><b><a href="/t/show-desktop-notifications-when-available/10183?page=4" itemprop="name" rel="next">next page ?</a></b></span>

Hey look! We are!

Hey look! There is!

Granted, there's not anything in <head> about this, and there could be, that's the single legit thing brought up in this topic... but isn't that overkill? Shouldn't proper rel=next and rel=prev links suffice?

I really do not want to add "PAGE X OF Y" to every <title> tag for these pages. It's meaningless, and far less important to the page than the actual title. So if we have a natural title of

Webmaster is giving errors of Duplicate Meta Descriptions and Duplicate Title Tags

then we "have to" add page numbers to the title because "the titles must be different" so...

Webmaster is giving errors of Duplicate Meta Descriptions and Duplicate Title Tags - PAGE 1 OF 3


As for your other concern about users editing titles and creating 5 "duplicate" titles for a topic -- irrelevant. A canonical link is on every page, and it always points to the current name of the topic.

<link href="https://meta.discourse.org/t/show-desktop-notifications-when-available/10183?page=3" rel="canonical" />

So rather than bothering us about this, @mohammad, when the data was all on the page for you to look at just as I did... why don't you ask Google why

  • their own Webmaster Tool doesn't respect the proper canonical links that we put in every page?

  • the properly formatted, as per their own documentation, rel=next and rel=prev links on the pages, don't allow them to detect pagination?

Because the evidence I have in front of me says we're doing everything exactly as documented by Google. But don't take my word for it! Feel free to take a look yourself, change your user agent to googlebot, load a longer topic, and view source yourself.


Sorry was busy just back to work.

@codinghorror How would you respond to this?

Every page is given its own canonical then how would the robot know which is the preferred URL?

These links are treated as different sets of links by Google and webmasters treats it as Crawling error.
Shall I ignore these errors or they pose a serious threat? Aren't you getting the same errors for Meta Discourse?

I agree on this but Who would tell that to Google?


Yeah, shouldn't the canonical links be the /t/slug/id/post urls?

Hmm, can we specify a URL for each <article>?


Was there an outcome to this thread? I am seeing the same duplicate meta descriptions and title tags error in Webmaster and want to know if I need to change something or whether it might be a false negative.

For example one of the errors I am seeing with duplicate meta description and title is on the urls: /206, /206/25, /206?page=2. In my …/206?page=2 canonical it is referencing itself, should it reference the page /206 instead to prevent the duplication?

Would one option be to edit the URL parameters in Webmaster to define how it crawls the pagination?