EU Label joint degree and multiple issuers for the same VC

[paulojorgeqferreira]

The new EU label initiative for joint masters and the European Universities alliance's ability to have a legal status raise the need for multiple issuers for the same VC. Open Certs support multiple issuers. Will the VC data model support multiple issuers ? Or VC presentations will cover this allowing several credentials to be mixed for a Joint degree? But for the EU label credentials maybe this option will not be the best one since it should be one VC only. Thanks

[RieksJ]

I strongly oppose the idea of VCDM to support multiple issuers for a single VC.

First, it may cause semantics problems for the contents. Semantics are not schemas, but the interpretation of schemas, and each party gets to interpret any schema in its own way. Having multiple issuers on a VC ensures that a verifier no longer has an authoritative source for the interpretation of the schema that was used, and hence also no authoritative source for the meaning of the claims. It also allows every issuer to refer to the others in cases where responsibility/accountability of the claims is discussed.

Second, I think it will cause protocol problems: how does one know that a VC has been signed by every issuer? What does a practical protocol look like that ensures this, given that every signature is likely to originate from specific parties each of which has its own way of doing things?

And there are good alternatives, the easiest of which is perhaps to have one party issue the VC, and other parties issue (VCs that hold) endorsement claims for that VC. There will be other alternatives.

[paulojorgeqferreira]

Thanks RieksJ. Can you provide some additional information about the endorsement claims for a particular VC. How would it work in terms of data? Each claim will in some way endorse a VC issued by a third party (referring to a did?), or it will be at VC level? Sorry to ask so many questions. I'm not an expert on the subject.

[dlongley]

+1 for first trying to solve new use cases through the composition of well-known, well-understood existing primitives -- before deciding to invent new ones.

[RieksJ]

The only work that I know exists is a paper that's being constructed as part of RWoT 11. If you're interested, you could decide to contact any of its authors, e.g. @dmitrizagidulin.

[vongohren]

Just throwing out some thoughts that has been roaming the community.

Group the issuers under a collaborative DID.

Not any work that I know of. Been discussed in conversations

Multi signatures

An email discussing this: https://share.streak.com/4IYzN5hJAU90r5AUIjrkx9
Uses chaining and represented in data integrity: w3c/vc-data-integrity#26

Other issues discussing multi issuer problem

#932
#741

[paulojorgeqferreira]

I did some additional research trying to find use cases and found additional approaches to the problem. I'm not familiar with the dynamics of w3c working groups and the relationship with the official recommendations. From the actual VCs data model, the suggestions below seem not aligned with One VC = One Issuer paradigm.

• From the Modeling Educational Verifiable Credential draft report (2021) there's a description of the use case of joint issuers.

• From the Verifiable Credentials for Education, Employment, and Achievement Use Cases (2022) there's a description of a use case about the transcript of records being Signed by More than One University - "A student is issued a transcript of the courses she’s taken at various universities. Each university has signed the transcript."

From the discussion here, it seems that endorsement claims, collaborative DIDs or Multi signatures are possible ways to achieve the need to have a VC that can be issued and verified proving that the holder was a student in a joint program. I will keep researching this (please share any additional resources that you may find.)

[RieksJ]

@paulojorgeqferreira: while the first text claims a need for multiple issuers, their only argument is that a single issuer is being overly restrictive in some use-cases. The second text talks about some transcript being signed by multiple universities, which is not the same as needing VCs to have multiple issuers (you can put the transcript and its signatures in a single claim within a regular VC).

@vongohren mentions texts that argue that there are use-cases that benefit from multiple parties signing data (e.g., contracts), and list-messages that go into implementation details, but no compelling argument for changing the number of issuers in VCDM.

[Sakurann]

Duplicate of #932. closing this one in favor of consolidating the discussion.