Page URL = canonical link

[Webliberty]

The URL of the page in the search bar of the browser must match the canonical link. If the canonical link does not contain the parameter f, then the URL of the page must also be without f.

The URL of the page in the search bar of the browser:

Code: Select all

viewtopic.php?f=1&t=1

The canonical link:

Code: Select all

viewtopic.php?t=1

Forum users copy the page addresses and share them, for example, on social networks or on other sites. Thus, all links do not lead to canonical pages, which is bad for SEO since search robots have to crawl 2 times more pages, creating a load on the server.

[Webliberty]

Why not do it in phpBB 4.0?

[Webliberty]

I just found information that removing the parameter f from the URL is already being considered and there is a corresponding ticket https://tracker.phpbb.com/browse/PHPBB3-16138

And also on GitHub: https://github.com/phpbb/phpbb/pull/5760

It is very cool!

[AmigoJack]

the search bar of the browser

Why isn't that the address bar to you?

[Webliberty]

Why isn't that the address bar to you?

In modern browsers, the search bar and address bar are combined and essentially the same.

[phponly]

So this means only the index of the topic will be indexed correct?

E: Thanks for helping me understand better guys.

[Brf]

As I understand it, the idea is to index each page of a topic only once. Each individual post should not be indexed, because a post's URL leads to its topic page, and would result in duplicate content for each post on the page.

The concern of this topic is that the canonical link does not contain the "f=" parameter, so does not match the URL of the page.

[Peter77sx]

bad for SEO since search robots have to crawl 2 times more pages, creating a load on the server.

You're right. There is no point for bots to crawl the search.php other than use up resources. The search page does not contain content to be indexed, it has links to follow, but that's kind of kind of the point of RSS feeds, you are able to submit RSS feeds to get crawled, and letting bots crawl your forums instead.

[KaileyT]

viewforum.php

Code: Select all

'U_CANONICAL'   => generate_board_url() . '/' . append_sid("viewforum.$phpEx", "f=$forum_id" . (($start) ? "&start=$start" : ''), true, ''),

viewtopic.php

Code: Select all

'U_CANONICAL'   => generate_board_url() . '/' . append_sid("viewtopic.$phpEx", "t=$topic_id" . (($start) ? "&start=$start" : ''), true, ''),

It should be easy to just add the f parameter to viewtopic.

[KYPREO]

bad for SEO since search robots have to crawl 2 times more pages, creating a load on the server.

You're right. There is no point for bots to crawl the search.php other than use up resources. The search page does not contain content to be indexed, it has links to follow, but that's kind of kind of the point of RSS feeds, you are able to submit RSS feeds to get crawled, and letting bots crawl your forums instead.

You have misunderstood the OP. He is not talking about search.php at all. He is talking about the browser navigation bar. Bots cannot and do not conduct keyword searches using search.php to identify topics - they crawl organically by following index, breadcrumbs, internal links etc.

The OP is right in that the canonical link for topics is viewtopic.php?t=*. This is how Google wants to index topic pages. However, when you (or indeed crawlers) navigate a phpBB board natively, viewtopic also contains the f=* parameter. This causes bots to crawl multiple versions of the same page. This can be verified in Google Search Console. I have over 100,000 duplicate ignored pages marked as duplicate of canonical.

I don't know that this necessarily "bad for SEO" as Google is smart enough to recognise the f=* version as a duplicate of the canonical. However, every page is getting crawled twice as often as it needs to be by every bot, placing unnecessary load on the server.

The f=* parameter is redundant as topic identification does not require forum categorisarion in the URL.

In any event, this change will happen as it has already been marked for merger into phpBB v.3.3.1.