Bug in the context file?

[iherman]

Consider a skeleton VP content:

{
  "@context": [
    "https://www.w3.org/ns/credentials/v2"
  ],
  "type": "VerifiablePresentation",
  "verifiableCredential": [{
    "@context": [
      "https://www.w3.org/ns/credentials/v2"
    ],
    "id": "http://university.example/credentials/1872",
    "type": "VerifiableCredential",
    "issuer": "https://university.example/issuers/565049"
  }]
}

per specification, the N-quads equivalent should be something like (I use CURIEs to make it more readable):

_:vp rdf:type cred:VerifiablePresentation .
_:vp cred:verifiableCredential _:credGraph .
<http://university.example/credentials/1872> rdf:type cred:VerifiableCredential _:credGraph .
<http://university.example/credentials/1872> cred:issuer <https://university.example/issuers/565049>  _:credGraph.

However, what I get on the playground:

<http://university.example/credentials/1872> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.w3.org/2018/credentials#VerifiableCredential> .
<http://university.example/credentials/1872> <https://www.w3.org/2018/credentials#issuer> <https://university.example/issuers/565049> .
_:b0 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.w3.org/2018/credentials#VerifiablePresentation> .
_:b0 <https://www.w3.org/2018/credentials#verifiableCredential> <http://university.example/credentials/1872> .

i.e., the "enclosing" graph does not appear.

This is either a bug in the @context file or the playground. If the former, it is the WG's job to solve it...

Do I miss something?

@dlongley @BigBlueHat @davidlehn @pchampin

[BigBlueHat]

@iherman fwiw, @gkellogg's Distiller creates these n-quads from the same JSON-LD...so there may indeed be a bug in the playground.

Here are the n-quads from the Distiller--which match what you were after:

_:g524580 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.w3.org/2018/credentials#VerifiablePresentation> .
_:g524580 <https://www.w3.org/2018/credentials#verifiableCredential> _:g524600 .
<http://university.example/credentials/1872> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.w3.org/2018/credentials#VerifiableCredential> _:g524600 .
<http://university.example/credentials/1872> <https://www.w3.org/2018/credentials#issuer> <https://university.example/issuers/565049> _:g524600 .

I'll do some more direct testing with jsonld.js (which is used on the Playground), and see how deep the rabbit hole goes. @davidlehn may have other thoughts.

[iherman]

@iherman fwiw, @gkellogg's Distiller creates these n-quads from the same JSON-LD...so there may indeed be a bug in the playground.

Here are the n-quads from the Distiller--which match what you were after:

_:g524580 <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.w3.org/2018/credentials#VerifiablePresentation> .
_:g524580 <https://www.w3.org/2018/credentials#verifiableCredential> _:g524600 .
<http://university.example/credentials/1872> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.w3.org/2018/credentials#VerifiableCredential> _:g524600 .
<http://university.example/credentials/1872> <https://www.w3.org/2018/credentials#issuer> <https://university.example/issuers/565049> _:g524600 .

I'll do some more direct testing with jsonld.js (which is used on the Playground), and see how deep the rabbit hole goes. @davidlehn may have other thoughts.

Phew. From the point of view of VCDM I am relieved (and also from the point of view of my own understanding...).

@msporny @brentzundel I am not sure whether we want to close this issue (being irrelevant for the VCDM spec) or leave it open until is handled, to avoid errors in discussions like in #1358 (comment) or #1358 (comment). I leave this decision to you. In any case, I removed the before-CR label.

[gkellogg]

I suspect it's a bug in handling scoped context in jsonld.js. Looking at the credentials/v2 context definition:

• VerifiablePresentation is a term which is used as a type in the outermost object, thus the type-scoped context it defines is used for expanding the rest of that object, which includes the term definition for verifiableCredential
• verfiableCredential which is in scope has a @container: @graph, meaning that its value is taken to be a named graph, which at least in my implementation, gets a blank node graph label. It also sets @context: null, but establishes a new @context which is used to expand the values of that graph object.

For some reason, jsonld.js does not seem to interpret the value of verifiableCredential as being a blank graph. I would have thought that some test would catch this case, though.

[pchampin]

FWIW, Sophia, which is based on @timothee-haudebourg's JSON-LD library, also parses this correctly: example.

[davidlehn]

There was missing JSON-LD test coverage for this use case and a jsonld.js bug. A JSON-LD test suite test was submitted, jsonld.js patched, and the json-ld.org playground was updated. It should work as expected now.

[iherman]

The issue was discussed in a meeting on 2023-12-06

• no resolutions were taken

View the transcript

2.8. Bug in the context file? (issue vc-data-model#1373)

See github issue vc-data-model#1373.

Brent Zundel: I think this one is before CR.

Manu Sporny: It's before CR.

Dave Longley: +1 to before CR.

Orie Steele: I would like some elaboration to be in the minutes: could someone explain the essence of the bug?

Dave Longley: It's not that we have a bug in our context file. It's that we added terms to make v1 VCs compatible with v2 Verifiable Presentations.
… We needed to make sure that context files for v1 VCs didn't affect the context files for v2 VPs.
… Some implementations had bugs, some didn't.
… We should be able to resolve this by nullifying the context as part of these implementations.
… We don't need to do anything in this WG.

Orie Steele: Sounds like there is no bug in our context, and this issue should be closed. It also sounds like there are buggy JSON-LD processors, that produce inconsistencies, that cause credentials to not verify.

Dave Longley: Orie: It is a reasonable expectation for any software is that there will be bugs from time to time.

Orie Steele: the complicated the software, the more reasonable it is to expect bugs.

Dave Longley: Orie: sometimes! :) ... i would say that it's not a good idea to assume simple software is less likely to be buggy (as this may lead to more likely bugs in such software).

Manu Sporny: You're coming across as trolling, Orie. Please stop. You asked us to explain something. We did, and now you're expressing what comes across as mock surprise.

[iherman]

The bug has indeed been handled and the example works as expected.

This means that the issue turned out not to be in our @context file, so there is no action for this WG.