Search engine optimization

[tobyzerner]

Flarum renders very basic versions of the discussion list and discussions (within <noscript> tags), which are going to be specifically optimised for search engines. This all takes place in Flarum\Forum\IndexAction and DiscussionAction, which extend Support\ClientAction and use Flarum\Support\ClientView. The actions get data from the Flarum's API and pass it onto content views (discussion.blade.php and index.blade.php). I'm happy to explain more if need be because it can get pretty confusing.

Put simply, the code is a mess. It needs to be rethought and refactored. A few thoughts:

• We probably want to use or build some kind of library that makes it easier to work with a JSON-API document payload.
• Similarly, we probably want to use or build some kind of library that will help generate URLs/query strings/etc.

Only after doing this will we be able to effectively optimise the SEO content and do the following:

• Add tags
• Add description tags
• Think about other stuff like schema.org
• https://moz.com/blog/seo-cheat-sheet

[axisdos]

Noindex for subpages (Inspired by Yoast SEO for WordPress) to avoid duplicate content. and URL Friendly.

[franzliedke]

@axisdos Noindex isn't needed when we have the rel=canonical tag. :)

[dcsjapan]

Questions of how to handle foreign language/unicode URLs -- such as #194 -- will also have SEO implications. In that regard, unfortunately, I think different regions may prefer different solutions.

[neandrake]

Would this tie into Accessibility at all? I'm not extremely familiar with it, but when structuring the noscript HTML along with the rest of the page, it would be nice if parts of it were identifiable for users with disabilities.

http://www.w3.org/WAI/intro/aria

[franzliedke]

Could this approach be an alternative to the noscript tags?

[tobyzerner]

Quite possibly, feels a little bit hacky though? I'm not too fussed

[dcsjapan]

#906 is an SEO-related issue specific to the Tags extension.

(I hope I was right to file it as a separate issue.)

[ghost]

I investigated links on my Flarum and posted results in this topic on official Flarum forum, have a look:
https://discuss.flarum.org/d/3103-is-the-flarum-s-seo-seriously-messed-up-or-am-i-doing-something-wrong

[franzliedke]

For reference: https://webmasters.googleblog.com/2014/02/infinite-scroll-search-friendly.html

[franzliedke]

Also interesting, with regards to changing <title> and other meta tags when doing client-side navigation: http://searchengineland.com/tested-googlebot-crawls-javascript-heres-learned-220157

[gwillem]

@01010000101001100 refers to a bug in the pagination code. Version 0.1.0-beta.7 will infinitely link to the "next" page id, even if there are no more discussions. I have only 1 page of discussions, but Google has already crawled to page 584. Fixing this would probably improve SEO a lot.

(--> #1310)

[franzliedke]

@gwillem @01010000101001100 Thanks for the heads-up. I've fixed the incorrect generation of pagination links, along with some other minor pagination-related improvements.

[askvortsov1]

To clarify, what exactly needs to be done here? Based off of the link Franz sent, Google seems to be fully capable of handling our frontend. Is it just a question of ensuring pagination is properly done?

I've been resolving meta title and description issues through a series of PRs, see #495 (comment), #2049

[BartVB]

Google is able to render Javascript but this is much, much slower than indexing 'static' HTML. More information here:
https://ja.dev/entry/blog/nagayama/render-budget-en

TLDR; use Javascript for endusers, HTML for crawlers.

[franzliedke]

Also, Google isn't the only search engine, especially in some other countries.

[franzliedke]

You're right, though, that we need to spec out more clearly what aspects we want to tackle. "SEO" is a very generic and not very actionable issue title. 😆

[xtrmus]

There are too many dofollow links on the site. ("Mentions, Likes, Replies") - Their attributes must be "rel = nofollow".
This item is very, very important. The reason for this: When Google bots enter the page, they see too many dofollow links and try to index them.

the ?near tag.
Google indexes both the near tag and the normal page. Therefore, because the two urls are trying to index in the same way, the bot is slow and cannot navigate the site in detail.

[askvortsov1]

Splitting into #2397, https://github.com/flarum/core/issues/626, #2397