Specify relationship between vct in sd-jwt vc and type in ld payload

[tlodderstedt]

The relationship between ' vct' and 'type' is currently unspecified. I think it makes sense to specify that the 'vct' value should be one of the values in the 'type' array. In my observation, there is typically one main type in this array, the respective value could be used in vct.

[tlodderstedt]

Suggested text: "vct and type coexist, but vct MUST be set to one value in the type array."

[Sakurann]

I don't agree with the proposal. I think we thought about is and said that sd-jwt vc must define it's own vct that governs the payload outside "ld" claim and the presence of the "ld" claim; and type value governs claims inside "ld" claim. so i don't see the need for vct to be one of the type values. and current situation is ok IMO.

[tplooker]

I don't agree with the proposal. I think we thought about is and said that sd-jwt vc must define it's own vct that governs the payload outside "ld" claim and the presence of the "ld" claim; and type value governs claims inside "ld" claim. so i don't see the need for vct to be one of the type values. and current situation is ok IMO.

While I get this theoretically, can you think of any usecases that would benefit from this flexibility? As I cant really and if thats the case I'd prefer Torstens proposal where we do force on of the types in the ld payload to be the vct. The risk of not doing this is that we make querying these credentials via DCQL much harder IMO.

[Sakurann]

The risk of not doing this is that we make querying these credentials via DCQL much harder IMO.

that's true. if we do this, i think we should also add a sentence that when sd-jwt vcld is used, all user claims must be obtained from the ld object. my comment is based on the assumption that there is a world where even in sd-jwt vcld, user claims can be obtained from sd-jwt vc payload, not necessarily ld claims, which I thought we agreed we do not want to prevent yet. but if we are ready to remove that option, that's fine (chair hat off)

also @javereec made a good point that allowing vct to be one of the type values potentially changes the processing rules, that all sd-jwt verifiers need to look if ld claim is present first and look into vct only after that.

[Sakurann]

Asia WG call discussion not blocking startign 60day review. this can also be added as an extension, it's not a breaking change

[Sakurann]

WG discussion:

• keep it unspecified closing in a week unless objections

[jogu]

closing as it's been 3 weeks since Kristina's "closing in a week unless objections" comment.