-
New Feature
-
Resolution: Unresolved
-
Minor
-
None
-
Future Dev
This is related to IDEA-7 but will break the idea down into more tasks
Proposing:
- A new page api where a page can declare metadata. This wouldn't be a raw meta tag it would be a conceptual set of metadata. Some such as the page title already exist, but there could be specific overrides so the social shared title is different to the html page title.
- Many pages like a course page are not visible to guests but the enrol page is which is where you get redirected. We need a way for a page to declare its own metadata even if it would get redirected to either a login page or enrol page, but in a way which makes it very explicit which data is shareable or not.
- A core api when rendering the page will gather all the metadata both inferred and explicit and render it in the various ways it is needed by various other sites. This isn't purely just opengraph is might also be schema.org, or other metadata schemes.
- One technical challenge that needs to be prototyped is how it works when a page is normally redirected. There is examples in the wild which conditionally don't redirect when they detect a metadata bot and serve a minimal page which just has the basic shareable metadata. This needs some careful security and usability and seo testing but it seems viable in other software like wordpress:
https://passwordprotectwp.com/docs/password-protect-open-graph-meta-tags/ - If this basic metadata page is show accidentally to a human then worse case it must show the redirect link it would have normally shown so they can follow along to the enrol page or the login page.
- Generic metadata bot detection is done using UA sniffing. Normally UA sniffing is considered a bad thing, but this applies to browsers. In the world of bots this is not a generally bad practice.
- Various pages will want to generate an image as part of the metadata, this will get deferred to a completely new api (MDL-78805)
