Orthogonality of format subspecs vs. feature subspecs

Hi all,

I’d like to hear your thoughts on how we should model the relationship between different kinds of subspecs. So far, we’ve mentioned the following subspecs, amongst others:

- WebID-Turtle
- WebID-JSONLD
- WebID-TLS
- WebID-OIDC

Clearly, these are not of the same kind. -Turtle and -JSONLD are “format” subspecs, meaning they define the serialization formats to be used for exchanging documents [1]. -TLS and -OIDC, however, are “feature” subspecs, meaning they define the content to be included in such documents in support of specific features. These two types of subspecs are orthogonal to one another. For example, an implementor might offer WebID-TLS over WebID-Turtle while another might offer WebID-TLS over WebID-JSONLD.

How should we deal with this orthogonality? Should we eliminate it by forcing “feature” subspecs to extend one or more “format" subspecs? Should we support this orthogonality? What should be the language used by implementors in describing their implementations?

Note that this is not an RFC thread as I do not believe that this specific issue has ever been discussed in the past. However, I kindly ask you to try and keep the discussion focused on the above and be forgiving of imprecisions related to other aspects of WebID.

Best,
J.

[1]: I’m being intentionally generic with the word “document” as that’s not within the scope of this thread.

Received on Monday, 27 November 2023 15:59:51 UTC