The discussion revolves around the lack of a canonical URL tag on the Discourse homepage. michaeld initiated the conversation, highlighting two issues with not having a canonical URL: duplicate search engine index entries when changing hostnames or domain names, and avoiding index entries with arbitrary parameters.
Mittineague suggested that setting a preferred domain in Google Search Console can achieve similar results, but michaeld argued that this is not a suitable solution for all search engines and does not address the issue of arbitrary parameters.
The conversation continued with codinghorror questioning the need for not redirecting everything when changing hostnames, and michaeld providing examples of scenarios where this would be desirable, such as soft launches of rebranding.
Gulshan_Kumar shared a personal experience of dealing with duplicate content issues due to the lack of a canonical URL on the homepage, and hawm pointed out that meaningless URLs, such as those with parameters, are being indexed by Google.
Long story short: it seems like the homepage doesn’t have a canonical URL tag… ?
That’s relevant when you change the hostname of the site and don’t want to redirect everything yet,
and in order to avoid search index entries like http://forum.domain.com?blahblahblah in case someone links to it that way.
I do not see the purpose of this request?
The only possible reason I can see for this regarding the “home” page is for “link and ranking signals” when
Your server is configured to serve the same content for the www subdomain or the http protocol.
But this can be more easily achieved on an individual basis without touching any Discourse code.
Set your preferred domain
Tell Google which version of your site’s URL you prefer for your domain:
It’s in my post.
Yes, that’s a solution for Google that can be used in case the software does not use a canonical URL.
But I see this as a last resort and a means to fix up an issue with a site. I would indeed require fixing it on an individual basis and it does not solve the ?blablah issue. And… you do know that Google is not the only search engine, right?
I see it as the first and only necessary approach and I seriously doubt its necessary at all.
Of course, but also I think being obsessed about micro-managing supposed SEO “tricks” is a major waste of time, energy and effort.
This is not about SEO tricks or being obsessed.
It’s about people changing their hostname or domain name and after three months they find out that search engines picked up that change perfectly except for their home page, which is now twice in the search engine index.
I’m a bit confused here since the necessity seems very clear to me. I see this topic has even been reclassified as a feature, while I’m convinced that the code path for the home page just forgets to call canonical_url …
This seems like such a bad idea, I’m not sure why anyone would want to encourage it?
Not encouraging, but supporting it would be desirable.
One specific example where this would be handy is to have a soft launch of a rebranding, where you want to communicate the new URL to search engines before communicating it to the customers.
(Real life example)
Another thing is that a third party can cause a search engine issue by causing duplicate content issues by linking to your home page with arbitrary parameters.
This is caught on every page of a Discourse installation except for the home page - because of the fact that the call to canonical_url is just omitted there. I would make a pull request if I was able to determine the exact code path for the home page…
@michaeld Hello!
Ya, this happens. I can confirm this. Another reason, why we may like to have canonical for the Homepage is the permalink /latest and / shows and index duplicate content.
Similarly, I have been trying hard to noindex via GTM but yet I’ve not been able to see effect.
I hope to noindex these common pages from first day of Discourse installation.
Una URL como /latest?exclude_category_ids%5B%5D=20&exclude_category_ids%5B%5D=19&no_definitions=true&page=1 está siendo indexada por Google, lo que parece carecer de sentido para el usuario.
