eclipse-ee4j
Make specifications more discoverable by search engines and AI
Ivar has researched a way that specification projects can increase the discoverability of their specifications. See this attachment for details.
[EXTERNAL] Fwd: Duplicate URL Issue.pdf
More instructions will be provided here.
Hey 👋🏻 🙂
about the canonical URL in the spec HTMLs ... since the content of the spec from version to version contains a lot of "duplicated content" it probably wouldn't help much if we just point to the spec page of the same version (I was looking at the patches applied to some of the projects so far).. instead if we could have a current version, e.g. https://jakarta.ee/specifications/bean-validation/current/jakarta-validation-spec and then all other versions pointing to the current version in their canonical link that would be better. For that though we'd have to set up the redirect and then make sure that once we release the new spec version the redirect from current to that new version is updated.
Edit: not a redirect per se... but more of a symlink, e.g. :
- link with explicit version # https://docs.hibernate.org/validator/9.1/reference/en-US/html_single/
- link with a "stable" version https://docs.hibernate.org/stable/validator/reference/en-US/html_single/
From doing a bit of reading, canonical is supposed to point to the "best" link for the page. It seems fairly clear that canonical shouldn't link to a page with different content.
There might be other things we could do to improve SEO, like putting the spec version in the title or potentially asking search engines not to index certain versions of the spec, but I don't think saying that the latest spec page is the canonical document for an old spec version is the right approach.
The approach outlined in this issue was what was decided by the Eclipse web team. If you have an issue with their approach, I recommend following up with them.
