The search engine treats any page as a document. The document must have a title and description in the meta tags. And the document itself must have a header (h1). This is not SEO, this is markup. All this should be unique for each page (document). The site consists of pages (documents).
We don’t have any of this at the moment.
It may make sense to follow their recommendations.
Styles for text headers and H1, …, H6 tags Well-designed text headers help the user to navigate through the page and save time. It is also easier for the robot to recognize the document structure. This is why you should put the headers in the tags according to the document hierarchy.
Source Yandex (just because it is well written there, in detail): Presenting information on a site https://yandex.com/support/webmaster/recommendations/presentation.html
Edit: ui/src/components/post-listing.tsx
{this.props.showBody && post.url ? (
<h1 className="mb-0 d-inline post">
<a
className="text-body"
href={post.url}
target="_blank"
rel="noopener nofollow"
title={post.url}
>
{post.name}
</a>
</h1>
) : (
<h5 className="mb-0 d-inline">
<Link
className="text-body"
to={`/post/${post.id}`}
title={i18n.t('comments')}
>
{post.name}
</Link>
</h5>
)}
Add main.css (from Bootstrap)
h1.post {
font-size: 1.171875rem;
}
There are many sites: reddit, dev.to, stack exchange, ruqqus, discourse etc… that follow the recommendations of search engines.
Example in the photo above.
In the example above, I added nofollow, but it may be worth considering only the H1. nofollow tags and meta description, I think this is a separate conversation.
#226
Since Lemmy is a single page web app, it loads most of its data, including the page title that would go in the H1, after the initial page load. Doing this would require an isomorphic app, which is very difficult to do, I’ve tried and failed several times at it with lemmy. Its really only important for SEO, so its secondary to a lot of other things I’m doing now.
That is SEO, and so is a document hierarchy.
These are better as github issues also. And would you do a pull request for the noopen nofollow thing? Thanks.
I looked. Yes, the task is not the same as I thought. Hmmm …