Define claims display description and claims path query

[danielfett]

Mainly fixes #266 based on https://vcstuff.github.io/sd-jwt-vc-types/danielfett/fix-claims/draft-fett-oauth-sd-jwt-vc-types.html#section-8

Also Fixes #272
Also Fixes #271

[mickrau]

After review in our team: In principle, we like the proposal and support it.
We left two comments regarding claims object display property and value_type.

[mickrau]

We would like to give display properties to an object of claims. The wallet could use this property when collapsing the object's claims in the UI. Is this possible and would the following be the right way?
If so, this should be part of an example.

                  {
                    "path": ["address"],
                    "display": [
                        {
                            "name": "Address",
                            "locale": "en-US"
                        },
                        {
                            "name": "Adresse",
                            "locale": "de-DE"
                        }
                    ],
                    "value_type": "object"
                },
                {"path": ["address", "street_address"]},

[sschulz-t]

+1 for this! I think supporting this is mandatory for having a clear and defined way to present this kind of data to the user. Let's use the example from your documentation and extend it with an arbitrary example:

... 
"address": {
    "street_address": "42 Market Street",
    "city": "Milliways",
    "postal_code": "12345"
},
...
"work_address":  {
   "city": "Berlin",
  ...
},

I would expect/imagine this to be displayed in the ui similar to this:

place of residence
+- street: 42 Market Street
+- city: Milliways
`- post code: 12345
....
place of work
+- city: Berlin
+...

[paulbastian]

Suggested change

	* `claims`: OPTIONAL. An array of claims description objects as defined in (#claims-description), excluding the `display` and `value_type` parameters. The `mandatory` parameter here is used by the Wallet to indicate to the Issuer that it only accepts Credential(s) issued with those claim(s).
	* `claims`: OPTIONAL. An array of claims path queries as defined in (#claims-path-query). The parameter is used by the Wallet to indicate to the Issuer that it only accepts Credential(s) issued containing exactly those claim(s).

[Sakurann]

I put a comment in issue #266, but after today's WG call where a concern that this is a very big breaking change was raised, we discussed this internally with Microsoft engineers. We do not currently have use-cases with the claims that have nested structure and array-based solution is too big of a breaking change since we just finished implementing OpenID4VCI based on the text that went to ID-1 voting. So our preference is actually defining the new property nested which may contain descriptions about nested objects (original approach 2).

As a compromise for the existing implementations, my suggestion would be fixing the text that becomes ID-1 with nested approach. and than after ID-1, incorporating the array approach defined in this PR.

[danielfett]

So our preference is actually defining the new property nested which may contain descriptions about nested objects (original approach 2).

Then we need a new PR for that.

As a compromise for the existing implementations, my suggestion would be fixing the text that becomes ID-1 with nested approach. and than after ID-1, incorporating the array approach defined in this PR.

That would mean two breaking changes.

[pmhsfelix]

+1

[pmhsfelix]

In the W3C VCDM 1.1, a credential can be represented by a JWT with the object following the data model inside a "vc" claim (https://www.w3.org/TR/vc-data-model/#example-usage-of-the-id-property). It presume that the root element of the Credential refers to the value of the "vc" claim and not to the JWT payload object, right? If so, should we be explicit about it?

[pmhsfelix]

I guess a similar question may be made when using SD-JWT with JWS JSON Serialization (https://www.ietf.org/archive/id/draft-ietf-oauth-selective-disclosure-jwt-14.html#section-8) - the root element of the Credential is the JWS payload object and not the top-level JSON object of the serialized credential, right?

[danielfett]

I'll change the wording to "Select the root element of the Credential, i.e., the top-level object of the
JSON structure representing the claims in the Credential"

[paulbastian]

I would prefer to exclude the value_type, as it seems under-specified and is not present in any example. Otherwise, I'd be ready to approve.

[danielfett]

I would prefer to exclude the value_type, as it seems under-specified and is not present in any example. Otherwise, I'd be ready to approve.

Done

[selfissued]

Looks good. Please answer my questions on particular lines before merging.

[awoie]

I still don't like the path syntax for ISO mdocs but @leecam convinced me that it would actually save some code.

[c2bo]

Let's move the other discussion into separate issues - definitely a big improvement to align with the dcql syntax

[sloops77]

agreed to move outstanding discussions to other tickets