Change typ values for JSON-LD claim sets?

[OR13]

currently we have:

vc+ld+jwt
vc+ld+cwt

I think this may not be correct, since +ld is not a valid suffix?

However I like this approach because of how intuitive it is compared to

credential+ld+json

potential alternative would be:

vc-ld+jwt
vc-ld+cwt

[TallTed]

I strongly advise discussing media types with their full names, rather than dropping the application/ that JWTs advise to omit when building the JWT (I could not find similar advice about CWTs), but which is then treated as being invisibly present.

I think this may not be correct, since +ld is not a valid suffix?

The application/ suffixes which must be valid do not include +ld; rather, they are
+ld+jwt (and +jwt) and
+ld+cwt (and +cwt),
which means that
application/ld+jwt,
application/jwt,
application/ld+cwt, and
application/cwt
must be defined and registered, before or simultaneously with definition and registration of
application/vc+ld+jwt and
application/vc+ld+cwt.

If I correctly understand the interrelationships of these types, a better set might be
application/vc+ld+json+jwt,
application/ld+json+jwt,
application/json+jwt,
application/jwt,
application/ld+cbor+cwt,
application/vc+ld+cbor+cwt,
application/cbor+cwt, and
application/cwt,
because these are somewhat likely to increase comprehension of derivation,
though I realize cbor and json are implicit in cwt and jwt, respectively,
and neither application/json+jwt nor application/cbor+cwt are really subtypes of application/jwt or application/cwt ....

All that said — I am not convinced of

• the intuitiveness of any of these to people outside our groups
• that we have decided how to handle {presentation|verifiable presentation} media types alongside {credential|verifiable credential} media types
• whether we have decided to have a single media type that encompasses both verifiable (with proofs) and non-verifiable (without proofs) {credentials|presentations}, and if so, whether that/those type/s reflects the possibility of verifiability

[thomas-fossati]

potential alternative would be:

vc-ld+jwt
vc-ld+cwt

The +cwt structured suffix is not present in the SSS registry. In an IETF-RATS draft we are requesting its registration though, so you may want to keep an eye on that document -- and possibly help pushing it through the RFC Editor line ;-)

[OR13]

I don't think we can call JSON-LD secured with a CoseSign1 a CWT, based on this:

https://mailarchive.ietf.org/arch/msg/cose/sTWvCwbX6RjMi1pV3kTn6huX1fk/

[OR13]

This PR can be closed when #120 is merged

[Sakurann]

PR was merged