- From: Lars G. Svensson via GitHub <sysbot+gh@w3.org>
- Date: Thu, 16 May 2019 10:32:47 +0000
- To: public-dxwg-wg@w3.org
Thinking about this again, I really think we need to publish the cartesian product of media type and profile URI. If we decide to make it a two-step process it would (in my understanding) look like this: 1. Client asks for a resource expressing interest in one or more profiles and one or more media types 2. Server answers with some content (according to its content negotiation algorithm) and also adds a link header that says in which profiles the content is available (but not which combinations of media type and profile) 3. Client wants to figure out if there is another combination of profile and media type that could work as well and sends a new request using one of the profiles listed in the link header and one or more media types it could consider. 4. Server answers `406 Not acceptable` and says "No representation available for this combination of profile and media type" 5. Client tries again with other combination (but sort of stabs in the dark since it has no clue which combinations are valid) 6. Server answers `406` 7. Etc... An alternative could be that if a client _only_ sends an `Accept-Profile`-Header with only one value (and no `Accept`-Header) the server would include the all possible combinations of profile and media type _for that profile_ but that would make client and server implementation much more complex. -- GitHub Notification of comment by larsgsvensson Please view or discuss this issue at https://github.com/w3c/dxwg/issues/785#issuecomment-493013524 using your GitHub account
Received on Thursday, 16 May 2019 10:32:49 UTC