The W3C Credentials Community Group

Verifiable Claims and Digital Verification

Go Back

Credentials CG Telecon

Minutes for 2017-08-29

Ryan Grant is scribing.
Christopher Allen: Today is a joint meeting of W3C verifiable claims (members only), and credentials comminuty group.
Christopher Allen: David Chadwick presenting

Topic: Introductions

Paul Simmonds: Hi, I'm Paul Simmonds from the Global Identity Foundation. We want a safer world where all entities can interact using a single digital identity; that provides them enhanced security and privacy, and which is completely under their control; enabling an eco-system of assured trust for all digital transactions. There is a lot in common between our work and the work being done here.
Ed Bice: Hi Ed Bice from Meedan. Meedan builds digital tools for global journalism and translation. We are a team of designers, technologists and journalists who focus on open source investigation of digital media and crowdsourced translation of social media. We are based in San Francisco and are doing work on collaborative fact checking with our Check product. We have been dealing with the Fake News problem since before it was called that. We heard about this group through Evan Sandhaus of the NYT and would like to explore if we can use Verifiable Claims to address some of our use cases. Here to learn more.
Ed Bice: More on us here...

Topic: Work Item Progress

Christopher Allen: How are we keeping track of work items?
Kim Hamilton Duffy: DID spec received discussion, so let's cover that. Data minimzation may have other groups working in similar areas.
Nathan George: Mentions competing efforts at hyperledger
Nathan George: There, there is an attempt to understand coverage of privacy related issues?
Adam Lake: +1
Manu Sporny: Fake news has brought us Ed and good feedback from BBC.
Manu Sporny: Not sure where Ed's work will land, we need to get them scheduled between VCWG and CCG.
Nathan George: Hyperledger forum where most credentials and verifiable claims discussions are happening
Mike Lodder: I would like to participate in the selective disclosure group
Nathan George: The Architecture WG has also spawned a Privacy and Confidentiality sub-group that is addressing some of these topics
Christopher Allen: Suggests Ed review RWOT Fall conference agenda for relevance. Especially reputation systems and which parts of our current work are underlying infrastructure for that.
Kim Hamilton Duffy: Bots, Fake News, and VC discussion:
Nathan George: Additionally there is implementation work going on here (folks from this group will be around at IIW 25
Christopher Allen: Are we done with mission statement? (yes!)
Kim Hamilton Duffy: We need to notify W3C and some other groups of our changes.
Nathan George: We have been working on support for shared crypto libraries and tools for selective disclosure at Hyperledger, if you are interested in helping that effort, I'd love to talk to you about how to get more organizations involved there.
Manu Sporny: Manu has action item to follow up with Dan from EOS. they're going to participate in verifiable claims.

Topic: Privacy and Security Requirements

Kim Hamilton Duffy:
David Chadwick: Draft security and privacy requirements is here:
David Chadwick: Upon review, when the subject is the holder (versus when subject is not subject) there are some subtle differences.
David Chadwick: If the subject is a third party holder, does it depend on the subject as to whether the credential can be verified?
David Chadwick: Two scenarios
David Chadwick: Subject delegates credential to holder, allowing holder to access a website or whatever
David Chadwick: With negative credential, a service may accept a non-delegated negative credential.
David Chadwick: Tricky.
Manu Sporny: General comment - it does get tricky - on the document: it's straightforward, but Manu can think of lots of edge cases that contradict.
Manu Sporny: Where to take it next? explore use cases or refine for high-level understand?
David Chadwick: Intent is to work it into the lifecycle and data model documents.
David Chadwick: It is meant to remain high level. there are ambiguities when looking at details.
David Chadwick: We're going to spell out the cases more for security review. not sure if this is that document. maybe lifecycle doc.
Christopher Allen: Document title: "A terminology for talking about privacy by data minimization: Anonymity, Unlinkability, Undetectability, Unobservability, Pseudonymity, and Identity Management"
Christopher Allen: We don't need to adopt their wording. We should look at attacker motivation, i.e. how would availability or integrity fail.
Manu Sporny: Here is the final RFC:
David Chadwick: Great suggestion
Manu Sporny: See published RFC (6973) defining private considerations. we would set ourselves up for failure if we didn't build off this work.
*Privacy considerations
RFC6973 compliance is already required in DID specification section 10.1
David Chadwick: If we show how we're using their guidelines, that helps our review process
Mike Lodder: Is a cryptographer working on selective disclosure
Christopher Allen: Separate work item exists: survey of non-cryptographic techniques surrounding data minimization. Looking for survey of the class of problems.
Christopher Allen: As a culture, we have a tendency to always use cryptography, even when it doesn't add security.
Kim Hamilton Duffy: We need more people dedicated to data minimization. RFC is good catalyst.
Manu Sporny: This is also very important for verifiable claims working group. previous criticism said we weren't doing enough deep dives into the space of techniques.
Mike Lodder: Do we have a formal place to begin working on it
Manu Sporny: Work now will reduce objections in a years time, when we're closer to finalization..
Dave Longley: And this CG can submit a report to the WG with some of that security/privacy information (and/or it can take place in the WG directly depending on the charter)

Topic: DID Specification Progress

Christopher Allen: Now that we're implementing DID-methods, we've discovered some issues.
Kim Hamilton Duffy: Looking for plan as to how we'll be making more progress
Manu Sporny: There is a new DID-method spec
Manu Sporny: Veres One DID Method specification:
Manu Sporny: Live site is here:
Manu Sporny: Hope is that this implementation helps us in the standardization discussion.
Manu Sporny: What is the impact? BTCR DID-method spec work raised lots of queitons.
Manu Sporny: We now have three different data models as suggestions.
Manu Sporny: Ethereum and Sovrin input will go into the spec.
Manu Sporny: Hope is to have discussion in September
Manu Sporny: Then come to technical decisions at conference
Manu Sporny: Has put up "straw men" as discussion points on proofs versus signatures
Dave Longley: Use cases, use cases, use cases
Manu Sporny: We believe we have generalized, but need to check everyone's use cases.
Manu Sporny: What is minimum viable DDO?
Manu Sporny: Want some kind of proposal going into RWOT
Manu Sporny: Need to prioritize
Christopher Allen: Issue identified: which keys can be used for control versus update
Manu Sporny: We agree that that (<--what?) is what should be done
Mike Lodder: +1
Manu Sporny: Two use cases: authentication versus update
Manu Sporny: Use case identified: want (or was considered) to prevent key from updating certain fields of DDO
Manu Sporny: Be very specific about proof models
Dave Longley: We were looking at separating out authentication from authorization
Dave Longley: This is about what fields you can write to in the DDO
Dave Longley: And you can constrain them to say that they can only authenticate using certain methods
Manu Sporny: General read on the DID spec: conflation between authentication and authorization
Manu Sporny: Went back to review Joram use case
Manu Sporny: We were able to use data structures to walk through use case.
Manu Sporny: Want to present at RWOT
Christopher Allen: We need to get Ethereum people back here
Ryan Grant: What was the conflation? [scribe assist by Manu Sporny]
Manu Sporny: It was design flaw in logical intent that led to a confused deputy attack when delegating and then the total loss of identity (read: it was really bad)
Christopher Allen: The blockchain CG is shutting down. there were possibly a couple issues to import, that were being discussed there.
Christopher Allen: The problem is can Web Ledger work with multiple blockchains
Manu Sporny: There was a futile attempt to align the data model across all blockchains.
Manu Sporny: Web Ledger Protocol -
Manu Sporny: As you all know, blockchains are just state machines... you get in events, write them to the ledger, that updates the state machine. We've been working on a unified data model and protocol that would enable generalized blockchain clients (just basic ability to read blocks and events... the contents are always going to be specific to the blockchain). Think of it kinda like HTTP for blockchains.
Manu Sporny: It could be that this group picks up some of this
Manu Sporny: This group could pick up Veres One. this isn't a central work item for the group
Christopher Allen: Interested in general concept of the API.
Christopher Allen: Some work to untangle things.
Christopher Allen: Good to see things being registered. this group could be a placeholder for things so that they don't get lost.
Manu Sporny: Microsoft has released their own blockchain, the Coco Framework
Manu Sporny: It feels like Hyperledger Indy in design, any feedback from Sovrin/Evernym folks?
Manu Sporny: What is the interplay between that work and the DIF?
Christopher Allen: Drummond isn't here, anyone else from Evernym want to say something?
Dave Longley: Coco Framework
Manu Sporny: They say "it's a framework for blockchains", but it looks like a specific blockchain
Christopher Allen: Also confused between Sovrin and Indy, which is hyperledger's implementation
Moses Ma: Talking to Aeternym in Paris. They're working with ISO, and coming in October (to RWOT) to bridge the groups. we should work harder to collaborate.
Nathan George: Sovrin is an installed instance of Hyperledger Indy for global public identity. Just like you can run multiple instances of a database like MySQL a network can run an installation of a blockchain, Sovrin is an instance of Indy for public identity.
Christopher Allen: We want to make a presentation at W3C TPAC
Dan Burnett: VCWG chairs are working on agenda
Christopher Allen: Kimhd what's our plan?
Kim Hamilton Duffy: Request from Christopher Webber to cover his project, else DID work.
Manu Sporny: Let's talk about data model for DDOs.
Christopher Allen: We should invite Christian from Ethereum uPort. I'll do that.
Christopher Allen: Thx all. See everyone next week!