Skip to content

Latest commit

 

History

History
93 lines (61 loc) · 6.34 KB

001_proposal.md

File metadata and controls

93 lines (61 loc) · 6.34 KB

Linked Claims

Metadata

  • Start Date:
  • URI:

Summary

The intention is to have a broad standard that is implementation independent, for making claims that can be connected together across different platforms, storages and implementations. Such claims may be public or private/access-gated.

In order to connect claims together, they must be addressable. A key part of this work is to define what it means for a claim to be addressable and qualify as a LinkedClaim. That is, for a claim to have a long lived identifier so that it is possible to make validation or audit claims ABOUT the claim itself.

patterns enabled by LinkedClaims

The goal of this work is to finalize and release the high level spec for LinkedClaims with https://github.com/Cooperation-org/LinkedClaims/blob/main/LinkedClaimsRFC.md as a starting point, with input from potential users and implementors.

Motivation

Trust shapes every interaction in our digital lives, from verifying academic credentials to tracking humanitarian aid. Yet we lack a universal way to connect and verify claims across different systems. Verifiable Credentials enable the secure signing of a credential, but the credibility of that credential still relies on some trusted central registry, or cannot be verified at all.

LinkedClaims addresses this need by providing a minimal but powerful specification for addressable claims and the cryptographic linking of independently issued credentials or attestations, referred to here simply as "claims". This enables verifiers to evaluate the credibility of claims based on their own roots of trust, as well as including data from both within and outside the Verifiable Credential ecosystem. LinkedClaims also enable individuals and organizations the agency to make and verify claims on their own terms, including the possibility of adversarial claims. Whether tracking disaster relief funds, documenting community development, or validating project reviews and research papers, LinkedClaims creates chains of trust that can be verified and connected.

image

Further, LinkedClaims invite layering of trust over time, as discussed in the "Progressive Trust in Issuer Registries" document below. The intitial creation of a durable identity does not in itself create trust, but it enables the later addition of attestations.

The real nature of trust is decentralized, progressive, and relative to the trusting party. By facilitating rich shared datasets that inform trust decisions while satisfying domain specific requirements, we believe LinkedClaims will pave the way for a more connected and verifiable digital world.

Coordination/Correspondence

The intention is to hold both technical workshops and discovery with stakeholders who may be potential users. In particular existing technologies may be recognized as following the LinkedClaims requirements and used as demonstrations of linking together existing signed claims.

Prior art and References

Proposed RFC for Linked Claims

IIW39_S1E_Progressive Trust in Issuer Registries with LinkedClaims

Composing Credentials via LinkedClaims and Cryptographic Binding published in Rebooting the Web of Trust XI proceedings.

Live demo of a general implementation of LinkedClaims at https://live.linkedtrust.us

Live demo of a skills credentials implementation of LinkedClaims at https://opencreds.net sponsored by the US Chambers of Commerce Foundation and the T3 Innovation Network.

We have interest in the concept of LinkedClaims from the UN CRM/CEFACT committee on critical raw materials, the US Chamber of Commerce Foundation, SidiHub Digital Identity Community, and some other groups and individuals.

The intention is to have this be a lightweight open standard.

Unresolved Questions

  • Is there a need for a specific "glue" vocabulary such as http://cooperation.org/credentials/v1/
  • How to resolve the tension between using a human-viewable subject URI versus a hashable signed credential as the subject of a claim? Could there be a content type or query parameter that switches, but in this case how to ensure the two are aligned?
  • Should the canonical identifier of a claim be included in the claim itself?
  • Review each of the MUST, SHOULD and MAY in the RFC for usefulness and testability
  • Can we write tests for being a LinkedClaim without a single canonical vocabulary?

and the largest question

  • What would motivate stakeholders to publish LinkedClaims and expose their claims to external validation? Do non-protocol entities see a use for this?

Outputs