The use of the fetch algorithm for image resources is optional and underspecified

[youennf]

The use of the fetch algorithm for image resources is optional and underspecified:

1. The spec says: If the user agent wants to display an artwork image, it is RECOMMENDED to run the fetch image algorithm. This is observable to JS via service worker. Should we use a MUST instead? (Safari seems to always use fetch)
2. The fetch algorithm is not setting the destination, which should probably be image (Safari sets it from quick code inspection).
3. While the fetch algorithm is using no-cors, response type is default, which forbids third party artwork (Safari does not do this check from quick code inspection).

In addition to that, a few items could be looked at:

1. If one artwork loading fails, it is not clear whether further loading should be pursued (Safari does further loading from quick code inspection).
2. When loading is actually triggered is not very clear.
3. Web application is not made aware if all artwork loading fails. Should it?
   For these items, I guess we could look at the html image element behavior/spec.

@chrisn, this should cover most of the issues you mentioned at TPAC 2024.
Do you see additional points we should cover here or as separate issues?

@steimelchrome, @padenot, any spec or implementation feedback?

[chrisn]

The other issue was related to ChapterInformation, I'd like to investigate further before raising it as an issue.