change ZKP section

[brentzundel]

This PR:

• Adds a version of the note initially drafter by Orie in Add issue to zkp section #1026
• modernizes and generalizes the language in the section in an attempt to make it more applicable to various zero knowledge proof mechanisms.
• adds a note that the examples will either be re-worked or removed.

I believe this PR prepares this section for for one of these possibilities:

1. The WG normatively specifies a method for securing VCs that supports ZKPs, in which case the examples and normative language in this section could be easily updated.
2. The WG does not normatively specify a method for securing VCs that supports ZKPs, in which case the examples and normative language in this section would be removed, and the remaining language could be moved to the implementation guide.

Signed-off-by: Brent Zundel [email protected]

---

Preview | Diff

[iherman]

The issue was discussed in a meeting on 2023-02-08

• no resolutions were taken

View the transcript

3.2. change ZKP section (pr vc-data-model#1030)

See github pull request vc-data-model#1030.

Brent Zundel: should be straightforward to adjust further or get rid of the section.

Manu Sporny: orie suggests process in order. Doing that.
… processing #1026, #1028 then need review on #1030.

[dlongley]

I remain a bit concerned about over promising what ZKP can realistically do in various use cases, but this text hasn't changed that much from what was originally approved to be in the VCDM 1.1 spec and there will be another revision at some point in the future before 2.0 is finished, so +1. Thanks for cleaning this up!

[Sakurann]

Selective disclosure, derived claim and unlinkability are three distinct features. By ZKP do we mean a way to secure VCs that fulfill all three or at least one? This needs to be clear. In the former, I think it is highly unlikely we will have a mature way before VC-data-model v2 CR, if the latter, SD-JWT will fall under this umbrella, but then the section name should be renamed from "ZKP".

[brentzundel]

Selective disclosure, derived claim and unlinkability are three distinct features. By ZKP do we mean a way to secure VCs that fulfill all three or at least one? This needs to be clear. In the former, I think it is highly unlikely we will have a mature way before VC-data-model v2 CR, if the latter, SD-JWT will fall under this umbrella, but then the section name should be renamed from "ZKP".

I agree that if SD-JWT becomes a normative part of the spec, then this section should be expanded to include it.

[TallTed]

Suggested change

	mechanisms with <a>verifiable credentials</a> requires an <a>issuer</a> to
	secure the <a>verifiable credential</a> in a manner that supports these
	mechanisms with <a>verifiable credentials</a> requires the <a>issuer</a> to
	have secured the <a>verifiable credential</a> in a manner that supports these

[Sakurann]

it should be clear that the properties being discussed are mainly Selective Disclosure and unlinkability (via derivatives and identifier clinding), and one mechanism does not have to achieve all three at the same time.

[msporny]

Approving with the assumption that minor nits and other people's suggestions are integrated.

[brentzundel]

@Sakurann and @TallTed I have made changes to the PR to reflect your comments. Please re-review.

[TallTed]

Much better than it was.

[msporny]

Merging this is currently blocked by @Sakurann. I've requested an explicit re-review. @Sakurann, please re-review the changes that @brentzundel has made to this PR so we can either merge it, or make further changes to address your concerns.

[msporny]

Merging this is currently blocked by @Sakurann. I've requested an explicit re-review. @Sakurann, please re-review the changes that @brentzundel has made to this PR so we can either merge it, or make further changes to address your concerns.

Still waiting on this (understanding that @Sakurann was on travel last week and unable to get to a review here).

[Sakurann]

Despite the change in the title, the section is still all about "zero-knowledge proof", which is a specific term and would not include things like SD-JWT. But given there are caveats to rewrite/remove this section written in the text, will keep the reservations until when that PR comes in later.

[iherman]

The issue was discussed in a meeting on 2023-03-14

• no resolutions were taken

View the transcript

2.2. change ZKP section (pr vc-data-model#1030)

See github pull request vc-data-model#1030.

Manu Sporny: I think we're ready to merge this.
… There are some changes that Kristina and Ted have suggested, but I think those are good to go unless someone here objects.

Dave Longley: +1 to merge with whatever editorial fixes.

Brent Zundel: ok, I'll apply language and merge.

[David-Chadwick]

Have you considered the impact of these edits on section 5.4 (data schemas) as this has an example that concerns ZKPs. Are edits also needed to this section?

[msporny]

Have you considered the impact of these edits on section 5.4 (data schemas) as this has an example that concerns ZKPs. Are edits also needed to this section?

If they are, they'll need to go in another PR -- this one is ready to merge. Please raise an issue if you think this needs to be fixed, @David-Chadwick or @brentzundel.

[msporny]

Normative, multiple reviews, changes requested and made, no objections, merging.

[Sebastian-Elfors-IDnow]

Thanks for the re-write of the ZKP section. It's a great improvement.

For completeness, the term "unlinkability" could be introduced. It is related to the bullet about "blinded signatures", which could be extended as follows:

"Blinded signatures also allows for unlinkable proofs, which removes a common source of correlatable properties about a holder during multiple presentations to one or more verifiers."