Skip to main content
Version: Stage 5 - Deliverable Acceptance Check

Public Review Detailed Log

Introduction​

This document contains a detailed log of all of the public review comments received by the GRID Project during the public review period. The text here was extracted from the GitLab Issues log for the project. The GitLab issues log remains the controlling source. The GitLab issues can be found here: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues

Where a merge request number is references (e.g. "!67"), then the merge can be found in the GitLab repository in the merge list here: https://opensource.unicc.org/un/unece/uncefact/gtr/-/merge_requests

Note that the "Author" shown is the account used to enter the issue in the GitLab issue register. The originating author can be found in the issue title or body.

Issue #112: TCPSP - Jasmine Jaegyong Chang/ISO TC154 WG6 convenor, Project leader of ISO 19626-3​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/112

Original Comment:​

UNTP Public Review Submission

Subject: Proposed clarification — GRID eligibility for Trust Communication Platform Service Providers (TCPSPs) performing identity-issuance qualification screening

Submitted by: Jasmine Jaegyong Chang/ISO TC154 WG6 convenor, Project leader of ISO 19626-3 Processes, data elements and documents in commerce, industry and administration — Trusted communication platform for electronic documents Part 3: Blockchain-based implementation guideline

Related document: UNTP Identity Anchor / GRID (Global Registrar Information Directory) specification

Submission type: Public review feedback — request for clarification / proposed normative text


1. Summary

The current GRID (Global Registrar Information Directory) specification establishes a trust hierarchy in which the Root is UN and Registrars are government agencies or industry/professional associations. This hierarchy does not clearly address entities that perform a distinct function: screening an applicant's legal/entity qualification at the point identity or account issuance occurs, independent of certificate issuance. We refer to such entities as TTPID(Trust Third Party Identity Directory) that enrolls entities(such as sender or receiver) and communication servers, and validates the fact of exchanging messages as TCE(Trusted Communication Evidence) on Trusted Communication Platform (TCP).

This ambiguity creates a practical adoption barrier for standards, such as ISO 19626-3, that reference GRID as the authoritative trust-anchor listing for identity issuers in cross-border B2B electronic trade. We request clarification on whether — and under what conditions — a TCPSP(TCP Service Provider) may be eligible for GRID listing, including recognition of existing Qualified Trust Service Provider (QTSP) status where applicable.


2. Background

ISO 19626-3 (in DIS Draft phase) considers to adopt the UNTP Digital Identity Anchor (DIA) model as the basis for its issuerDID trust-verification mechanism for cross-border federation: an EntityID's issuing authority is identified via a did:web-based identifier and its legitimacy is verified by checking GRID registration.

In mapping this structure to real-world implementation, we identified the following gap:

  • GRID's stated hierarchy (UN Root → government/association Registrar) implicitly assumes that qualification-screening at issuance is a public-sector or association-based function.

  • This function is distinct from certificate issuance (the CA function), which already operates under separate, mature trust infrastructure and is not the subject of this submission.

  • A growing share of qualification screening at issuance — particularly for cross-border trade entity onboarding — is performed by commercial platform service providers (e.g., electronic trade portals verifying business-registration status against authoritative government registries), some of which already hold Qualified Trust Service Provider (QTSP) status under recognized regulatory frameworks such as eIDAS.

  • Without an explicit inclusion pathway, standards built on GRID (such as 19626-3) cannot represent this category of entity within the trust hierarchy on a principled basis.

This issue parallels a structural question already under discussion in other UNTP-adjacent standards work: how an entity that simultaneously acts as trust issuer and as a registrant requiring its own trust assurance should be structurally accommodated (a dual-role registration problem**).**


3. Specific Question for Clarification

Does the current or planned GRID eligibility criteria include a pathway for Trust Communication Platform(TCP) service providers— entities performing qualification screening at identity/account issuance, distinct from certificate issuance — to be listed as Registrars, either (a) directly, subject to verifiable screening-process evidence, or (b) via recognition of existing Qualified Trust Service Provider (QTSP) status in lieu of separate government endorsement?

If no such pathway currently exists, we respectfully request that one be considered, for the reasons outlined in Section 4.


4. Proposed Text Amendment (for consideration)

We propose the following addition to the GRID Registrar eligibility criteria (exact placement and wording to be adapted by UNTP editors to fit existing document structure):

Proposed new clause — GRID Registrar Eligibility: TCP Service Providers

In addition to government agencies and recognized industry/professional associations, a Trust Communication Platform Service Provider (TCPSP) — an entity that screens an applicant's legal/entity qualification at the point of identity or account issuance — may be eligible for listing as a GRID Registrar where either of the following conditions is met:

(a) Direct evidence pathway — The TCPSP which operates blockchain DLT based identity directory demonstrates a verifiable qualification-screening process, such as real-time or API-based cross-verification against an authoritative government registry (e.g., business registration, corporate registry) applicable in the relevant jurisdiction.

(b) Qualified recognition pathway — The TCPSP holds current, verifiable status as a qualified trust service under a recognized regulatory framework (e.g., eIDAS or an equivalent regime meeting the "substantially equivalent or higher" standard of UNCITRAL Model Law on the use and cross-border recognition of Identity Management and Trust Services (MLIT), Art. 25/26), which shall be deemed to satisfy the trust threshold otherwise required of a government agency or association Registrar, without requiring separate case-by-case endorsement.

Listing under either pathway shall be subject to periodic revalidation, and the specific screening process or QTSP qualification relied upon shall be recorded in the GRID entry for transparency purposes.


5. Requested Action

  1. Confirm whether TCPSP eligibility, or recognition of existing QTSP status, is already addressed elsewhere in UNTP documentation that we may have missed, and point us to the relevant section if so.

  2. If not currently addressed, consider the proposed clause (Section 4) or an equivalent mechanism during the current public review cycle (closing 2026-07-13) or in the subsequent Technical Working Group session.

  3. We are available to participate in TWG discussion on this topic once the group reconvenes, and can provide further detail on how ISO 19626-3 depends on this clarification.


This submission is made in the context of ISO 19626-3 development under ISO/TC 154, referenced also via ISO TC154 and UNECE’s ebMoU and long collaborative joint activities.

Discussion​

[2026-10-28] John Phillips: Thank you for your submission and for detailing the alignment with ISO 19626-3. We appreciate the collaboration with ISO TC154 WG6 and recognize the critical role that Trust Communication Platform Service Providers (TCPSPs) and QTSPs play in cross-border B2B electronic trade.

After review by the project team, we have determined that the proposed amendment, to include TCPSPs/QTSPs as "Registrars", cannot be accommodated. The core reason is that this would conflict with the architectural scope and definition of a Registrar within the GRID specification.

The details of our thinking are as follow:

  1. Clarification of GRID’s Scope (Authoritative Source vs. Service Provider) In the UNTP/GRID architecture, a "Registrar" is strictly defined as the entity holding the legal, statutory, or recognized administrative mandate to act as the primary source of truth for a given domain (e.g., a National Business Registry, a specific Ministry, or a mandated Customs Authority).

    Entities that perform qualification screening—even highly regulated ones like QTSPs under eIDAS—do so by verifying claims against those authoritative registries. They can perform a valuable commercial and social role but in the context of GRID definition of "authoritative" having a legal basis, they are consumers of authoritative data, not the originators of it.

  2. Decision on Proposed Text Because TCPSPs act at the service/intermediary layer rather than the authoritative root layer, adding a pathway for them to be listed as GRID Registrars would conflict with the directory's purpose. Therefore, the proposed text amendment (Section 4) will not be integrated into the specification.

  3. Next Steps and Collaboration We understand that TCPSPs and QTSPs can be important participants and service providers in the overall UNTP ecosystem. The architectural relationship we see is that a TCPSP could use GRID to discover the correct authoritative government registry to perform its real-time or API-based cross-verification (as you rightly note in your "Direct evidence pathway" example).

We would welcome the opportunity to discuss this architectural relationship further in upcoming TWG sessions, specifically focusing on how ISO 19626-3 can successfully consume GRID data to empower TCPSPs in their identity-issuance screening, without needing to list the platforms themselves within the directory.

[2026-10-29] ALINA NICA GALES: Thank you, Jasmine, for raising this important question...

From the perspective of the Global Trust Registry (GTR) project, I would support the distinction between an Authoritative Registrar and a trust service provider performing identity verification or qualification screening.

The fundamental criterion for GRID eligibility should remain the existence of a legally recognised mandate to maintain an authoritative register, rather than the technical reliability of the verification process or the regulatory qualification of the service provider.

In this respect, QTSP status under eIDAS provides important assurances concerning specific qualified trust services, but does not, in itself, confer authoritative registrar status or the legal competence associated with an official register.

I would, however, suggest two further clarifications.

1- First, the decisive criterion should be the legal mandate and responsibility for the relevant register, rather than exclusively the public or private nature of the organisation. This would preserve flexibility across different national legal systems while maintaining the authoritative character of GRID.

2- Second, the relationship between authoritative registrars and trust service providers could be further clarified in the implementation guidance, particularly regarding how TCPSPs may rely on GRID-listed authoritative sources for identity verification and subsequently issue their own service-level credentials or attestations.

Such an approach would support interoperability with ISO 19626-3 without conflating the legal authority of the underlying register with the technical or regulatory trustworthiness of the intermediary.

It may also be useful to explore whether existing UNTP mechanisms for non-authoritative identifier schemes offer an appropriate interoperability pathway in relevant cases.

This distinction between authoritative legal sources and the services built upon them is, in my view, essential to preserving the legal certainty and international scalability of the GTR framework.

Alina.


Issue #109: DFGT S-10 - Schema Versioning and Governance Strategy​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/109

Original Comment​

Document​

Participating Registrar Data Schema (registrar-v1.json) / Implementation Guidelines

Issue​

The formal schema is already published as registrar-v1.json but no versioning or backward-compatibility strategy appears in any document. The schema will need to evolve as GRID participation grows, as new capabilities are added, and as requirements mature. Without a defined versioning governance framework, the first schema update will disrupt existing national integrations without warning. This risk increases with each new national participant that goes live under the current schema.

Proposal​

Define and publish a versioning strategy covering: (i) version numbering convention (semantic versioning or equivalent); (ii) how new optional fields will be introduced without breaking existing parsers; (iii) how mandatory fields will be added with a defined minimum transition period; (iv) minimum deprecation notice before fields are removed; (v) how prior schema versions will remain resolvable for historical verification of records issued under those versions. This can be published as a short governance annex or as part of the GRID operating procedures.

Discussion​

[2026-09-30] John Phillips: This is an important observation and we need to make the changes to be ready for operation.

Tl;DR: The UNTP specification has already had to address this issue and has adopted a semver best practice approach. GRID should do the same.

Design rationale and proposed approach​

The GRID project should adopt a schema versioning and governance framework modelled directly after the UN Transparency Protocol (UNTP) and aligned with GRID’s UN/CEFACT Open Development Process (ODP) and pilot architecture.

This means that UNTP will manage schema and context evolution using semantic versioning (SemVer) embedded into immutable URL paths (e.g., [https://untp.unece.org/contexts/dia/v0.7.0/context.json](https://untp.unece.org/contexts/dia/v0.7.0/context.json)), guaranteeing that existing integrations remain valid against the specific version they declared without breaking when specs upgrade.

Key Design Decisions for GRID​

UNTP-Aligned SemVer with Immutable Path URIs

This mirrors UNTP’s URI-based schema hosting and decouples schema updates from pipeline deployments.

  • URL Structure & Immutability:

  • Canonical Versioned Path: [https://opensource.unicc.org/un/unece/uncefact/gtr/schemas/v1.0.0/registrar.json](https://opensource.unicc.org/un/unece/uncefact/gtr/schemas/v1.0.0/registrar.json)

  • Major Alias / Active Link: [https://opensource.unicc.org/un/unece/uncefact/gtr/schemas/v1/registrar.json](https://opensource.unicc.org/un/unece/uncefact/gtr/schemas/v1/registrar.json) pointing to the latest non-breaking release in the v1.x branch.

  • Semantic Versioning Rules (MAJOR.MINOR.PATCH):

  • PATCH (1.0.0 → 1.0.1): Backward-compatible bug fixes or description updates (e.g., clarifying field $comment or description without altering data constraints).

  • MINOR (1.0.0 → 1.1.0): Backward-compatible additions. Introducing new optional properties (e.g., registrar_notes or query_endpoint). Existing sovereign node parsers continue to validate successfully.

  • MAJOR (1.0.0 → 2.0.0): Breaking changes. Converting an optional property into a required array item, changing string regex patterns (e.g., UN_member_state or subdivision_code), or removing existing fields.

  • Payload Schema Pointer:

  • Sovereign node edge-normalized JSON files declare their target schema via the standard $schema key: "$schema": "[https://opensource.unicc.org/un/unece/uncefact/gtr/schemas/v1.0.0/registrar.json](https://opensource.unicc.org/un/unece/uncefact/gtr/schemas/v1.0.0/registrar.json)"

Implementation: Proposed Governance Strategy (Resolving Issue #109)​

To satisfy the specific requirements listed in Issue #109, publish a GRID Schema Governance Annex defining the following five rules:

Requirement MDUNTP & GRID Governance Implementation Strategy
(i) Version Numbering





Adopt SemVer (MAJOR.MINOR.PATCH). Pre-production pilot versions use 0.x.y (aligned with ODP Stage 3/4), while production stability begins at 1.0.0 (ODP Stage 5/6).


(ii) Optional Fields





New non-breaking fields MUST be defined as optional (omitted from the schema's "required" array). Edge node parsers and central harvesters (generate-grid.js) MUST use permissive validation (additionalProperties: true or soft evaluation) so new fields do not trigger build failures.


(iii) Mandatory Fields





Mandatory fields cannot be added directly to an active MAJOR version. They MUST first be introduced as optional fields in a MINOR release with a minimum 6-month transition period before being promoted to required in the next MAJOR release.


(iv) Deprecation Notice





Fields flagged for removal MUST be annotated with "deprecated": true in the JSON Schema for at least 6 months prior to a MAJOR version increment.


(v) Historical Resolution





All published schema versions (v1.0.0, v1.1.0, etc.) remain permanently hosted and immutable at their respective URI paths. Historical Verifiable Credentials (DIAs) minted by the central harvester retain their validation integrity over time.


  1. Rename Artifact: Rename registrar-v1.json to registrar-v1.0.0.json and move it to static/schemas/v1.0.0/registrar.json in the project repo.

  2. Update Edge Harvester: Update generate-grid.js to validate fetched sovereign node JSON against the specific schema URL declared in the node's payload $schema field.

  3. Publish Governance Annex: Add a Schema-Governance.mdx document in the Docusaurus docs/ directory detailing the SemVer rules, deprecation policies, and historical resolution guarantees.

[2026-10-01] John Phillips: I'm proceeding with adopting these changes as they are both necessary and non-controversial. The updates will be reflected in the main branch once all tests are complete at which point this issue will be closed.

[2026-10-01] John Phillips: Updates made and pipeline tests passed.

New site content published.

This issue will be closed.


Issue #108: DFGT S-09 - Delegation Model and Personal Data Privacy Conflict​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/108

Original Comment​

Document​

DIA Legal Meaning and Data Structure Para 2 / Implementation Guidelines generally

Issue​

Neither the DIA document nor the Implementation Guidelines addresses how an authorised representative, officer, employee or agent of a registered entity is identified, credentialled and linked to that entity's DIA. In practice, a legal entity always acts through a human being with some scope of authority for some period. Without a standard model for this relationship, implementers will build incompatible local solutions, or rely on the informal assumption that whoever presents the entity DID is authorised to act - which repeats the conflation identified in S-02 above. A related privacy conflict: Article V of the DIA document prohibits personal information in the DIA, but delegation necessarily involves identifying a named person. These two requirements are in direct conflict without a resolution.

Proposal​

Define a standard conceptual model for authority and delegation: (i) the entity-level DIA establishes the registered entity's identity; (ii) a separate, interoperable authority or delegation credential links a named individual or automated system to that entity, with defined scope, effective period, and revocation mechanism; (iii) the authority credential is verifiable independently of the entity DIA but resolves back to it. This need not be embedded within the DIA itself. The privacy conflict is resolved by this separation: the entity DIA contains no personal information, while the authority credential handles the delegation and can be subject to appropriate purpose limitation.

Discussion​

[2026-09-30] John Phillips:

This is an important point and has had some discussion in the project team and meetings. The conclusion of those discussions was

  1. The GRID itself will follow privacy policy guidelines relevant for its jurisdiction of operation. The default expectation is that these will follow UN/CEFACT -> UNECE -> UN requirements.
  2. GRID participants decide what the GRID will show about their Registry's AND Registrar's. This means that the GRID will only present what has already been presented publicly by the GRID participant.
  3. GRID Participants have the sovereign responsibility for observing the privacy regulations of their jurisdiction. This means that what they decide to display in their Registry and issue in their DIA's (beyond the core UNTP information) is up to them. GRID has no authority to dictate how a Registry operates or what its jurisdiction deems private information.

This is addressed in the principles of the GRID design and in "Article 5: Data Protection and Privacy" of the DIA legal meaning and data structure document.

Conclusion No change proposed to existing documentation. When the GRID transitions to an operational status, THEN the Privacy regulations can be revisited with a concrete understanding of the jurisdiction under which the GRID will operate.

[2026-10-01] ALINA NICA GALES: The issue appears to conflate two distinct concepts: the authoritative identification of a registered entity and the legal authority of a natural person or system to act on behalf of that entity.

The purpose of the DIA is to provide a verifiable attestation of the registered entity's identity and its linkage to an Authoritative Registrar. It is not intended to establish or standardise representation, agency, powers of attorney or delegation.

Authority to act for a legal entity is jurisdiction-specific and may arise from different legal sources, including statutory office, registered representation, powers of attorney, contractual mandates or other forms of authorisation. These concepts cannot necessarily be reduced to a single delegation model.

Accordingly, control or presentation of an entity's DID or DIA should not, by itself, be interpreted as evidence that the person presenting it is authorised to act on behalf of the entity.

Where an implementation requires a separate authority or delegation credential, such a credential may reference the entity's DIA, but its legal meaning, scope, lifecycle and privacy requirements should remain outside the normative scope of the DIA and GRID specifications.

On that basis, there is no inherent privacy conflict within the DIA: personal data relating to representatives need not be included in the DIA at all.


Issue #107: DFGT S-08 - Clarification of Registrar-Provisioned Identifier Model​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/107

Original Comment​

Document​

DIA Legal Meaning and Data Structure Para 1.2 / Implementation Guidelines Paras 1.6, 1.7 and Section 3.9

Issue​

The DIA document Para 1.2 describes an applicant-controlled DID model: the applicant provides the DID and the Registrar verifies control. Implementation Guidelines Paras 1.6 (Technological Neutrality) and 1.7 (Proportionality and Inclusive Maturity) imply that other provisioning models are permissible but do not state this. No implementation guidance addresses the registrar-provisioned model, in which the Registrar generates, manages and controls the cryptographic identifier on behalf of the registrant. Implementers reading the current documents may conclude that applicant-managed keys are required.

Proposal​

Add an explicit statement to Section 3 that a registrar-provisioned cryptographic identifier is a permissible implementation of the DIA architecture, provided that a verifiable and trustworthy binding between the subject, the authoritative identifier and the cryptographic identifier is maintained. Extend Section 3.9 (Minimum Implementation Requirements) to include: the applicable control-verification process and its documentation for both the applicant-supplied and registrar-provisioned models. Reference Paras 1.6 and 1.7 as the governance basis for this flexibility.


Discussion​

[2026-09-30] John Phillips: The concerns regarding implementation friction and the need to support diverse technological maturity levels (Paras 1.6 and 1.7) are entirely valid. We agree that requiring all applicants to independently manage their own cryptographic keys can present a barrier to entry, particularly for SMEs, and that custodial solutions can accelerate widespread adoption.

However, resolving this by explicitly including a "registrar-provisioned and managed" model conflicts with the design principles of the UNTP Data Architecture.

The Design Implications of Registrar Custody​

The core purpose of the Digital Identity Anchor (DIA) is to decouple the authoritative trust anchor (the Registrar) from the daily transactional activities of the business.

In the UNTP applicant-controlled model, the Registrar verifies the entity, issues the DIA to the entity’s DID, and then steps out of the transaction flow. The entity is then free to use their cryptographic keys to sign Verifiable Credentials (VCs), Digital Product Passports (DPPs), and other trade claims across any platform.

If the specification standardizes a model where the Registrar generates, manages, and controls the cryptographic identifier on behalf of the registrant, it inadvertently shifts the Registrar’s role from an authoritative anchor to a centralized signing platform. This would have two consequences:

  1. Centralization of Trade Claims: To issue downstream VCs, the registrant would have to route their daily trade data back through the Registrar’s infrastructure for signing.

  2. Loss of Self-Sovereignty: This effectively creates a centralized platform architecture, contradicting the "protocol not platform" philosophy that underpins the UNTP, where trade entities must have independent cryptographic agency.

Separation of Concerns: Custodial Agents vs. Registrars​

The goal of inclusivity can be met without compromising the decentralized architecture by clearly distinguishing between "Applicant-Managed" keys and "Applicant-Controlled" keys.

An applicant does not need to manage their own keys locally; they can choose to delegate key management to a Custodial Wallet Provider or Software Agent. In this scenario, a third-party service provisions and manages the DID on behalf of the applicant, abstracting away the technical complexity.

Because the applicant legally controls their relationship with the wallet provider (and can theoretically port their identity or revoke the delegation), the DID remains functionally "Applicant-Controlled." Conflating the role of the Wallet Provider with the role of the Registrar is where the architectural conflict arises.

Proposed Alternative Resolution​

Rather than amending Section 3 to endorse registrar-provisioned identifiers, we propose amending the documentation to clarify how custodial key management fits within an applicant-controlled framework.

Proposed amendment to Section 3.9 (Minimum Implementation Requirements):​

"The DIA architecture mandates that the cryptographic identifier remains under the control of the applicant. However, 'applicant-controlled' does not strictly require local 'applicant-managed' infrastructure. Applicants may delegate the generation and daily management of their cryptographic keys to a designated Custodial Agent or Wallet Provider (which may include state-sponsored digital public infrastructure), provided that:

  1. The legal and operational boundary between the Authoritative Registrar and the Custodial Agent is clearly maintained.
  2. The applicant retains the right to revoke the custodial delegation or port their credentials to a different agent."

This approach recognises the requirement for frictionless, custodial onboarding while preserving the decentralized integrity and separation of powers required by the UNTP.

[2026-09-30] John Phillips: These changes have been made in the branch PRU03

[2026-10-01] ALINA NICA GALES: I think the issue starts from a wrong assumption about the purpose of the DIA.

The fact that a legal entity acts through natural persons does not mean that representation, powers of attorney or delegation must be included in, or evidenced by, the DIA.

The DIA is not intended to reproduce all authoritative information held by a registry. Representation and powers are separate legal matters, governed by national law and, in many jurisdictions, provided through specific registry information or certification services.

So, in my view, the absence of delegation data is not a gap in the DIA model. It is simply outside its scope.

I would only add a clarification that possession or presentation of a DIA does not, by itself, prove authority to act on behalf of the entity.

In short: the DIA is not a substitute for registry publicity, and not every authoritative fact held by a registry should become a DIA attribute.

Issue #106: DFGT S-07 - Normative Credential Lifecycle and Compromise Framework​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/106

Original Comment​

Document​

Implementation Guidelines Para 3.7

Issue​

Para 3.7 lists 'credential lifecycle handling' and 'mechanisms for status, revocation, expiry or freshness' as capabilities DIA-capable implementation should support, but does not define what lifecycle events must be handled or how. The following events are unaddressed: (a) periodic key rotation; (b) key compromise, emergency revocation, and recovery; (c) entity restructuring - how an existing DIA relates to a successor entity following a merger or demerger; (d) entity termination - what status a DIA carries after the underlying registration lapses; (e) preservation of historical verifiability after identifier or key changes. Each Registrar will handle these independently if no model is provided, creating inconsistent outcomes across jurisdictions.

Proposal​

Add a normative lifecycle model to Section 3.7 covering at minimum the five event types listed above, with required actions for the Registrar, required updates to the GRID record, required notification to affected credential holders, and required treatment of credentials already issued under the prior key or identifier. The model should address historical verifiability explicitly: a credential issued before a key change must remain verifiable through the issuer's publication of the prior verification material for a defined retention period.

Discussion​

[2026-09-30] John Phillips: This is an important topic and has received a significant amount of focus over the last few weeks and months, both within the UNTP community and more widely in the W3C Credentials Community. We are, we think nearing a resolution.

One comment on the issue(s) as raised is that the cryptographic key event management and the credential life cycle management are two separate, related and essential elements.

The design needs to address both areas.

[2026-10-01] Sankarshan Mukhopadhyay: I agree with John's observation of the distinction between cryptographic key events and credential lifecycle events. I would preserve that separation explicitly in the resulting model rather than treating them as one lifecycle.

A key can rotate or be compromised without the underlying registration or authority necessarily ceasing to exist. Conversely, a registration, mandate, delegation, or credential can expire or be withdrawn while the cryptographic key remains technically valid.

The model could therefore describe related but independent state transitions for:

  • verification material;
  • credential or attestation status;
  • underlying registration or mandate; and
  • where applicable, delegated authority.

The verifier then evaluates the relevant combination for the time and purpose of the transaction.

For historical verification, retaining old verification material is necessary in some designs but is not sufficient by itself. The verifier also needs to know the temporal effect of later lifecycle events. A key discovered as compromised in August, for example, may affect reliance on signatures from May if the authoritative determination says the compromise was effective from May.

This would possibly align #106 with #105 without making credential lifecycle and key lifecycle the same thing.

[2026-10-15] John Phillips: Proposed Resolution:

The concerns are relevant and valid however we must not over step the GRID's core tenet—that national Registrars operate under their own statutory mandates and must not be dictated to by the GRID.

Section 3.7 will be edited provide a normative baseline framework defining expected cryptographic and data lifecycle states, but allow jurisdictions flexibility in domestic workflow execution. The framework must mandate that prior verification material (public keys, DID log states) is retained indefinitely to ensure past signatures remain verifiable.

Updates will be made merge !68

Issue #105: DFGT S-06 - Temporal Verification and Historical Status Support​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/105

Original Comment​

Issue​

Both Para 2.7 and Para 3.7 address current status, revocation, expiry and freshness only. Neither addresses the ability to determine whether a credential was valid at a specified past date. Trade disputes, financing decisions, customs audits, and proceedings before courts and arbitral tribunals require point-in-time verification: was this entity registered and in good standing on the date of shipment, or on the date the credit facility was drawn? A system that can only attest current status is architecturally incomplete for these purposes.

Proposal​

Add to Para 3.7 (Technical Implementation Considerations for DIAs) a requirement to support temporal verification including: (i) issuer-signed timestamps enabling point-in-time proof of credential validity at the time of issuance; (ii) a status log mechanism or equivalent that preserves the state of a credential at defined historical points; (iii) verifier guidance on how historical status should be determined and presented for reliance purposes. Timestamping per RFC 3161 or an equivalent standard provides an established technical basis.

Discussion​

[2026-09-30] John Phillips: We expect to resolve this issue at the same time as #106

[2026-10-15] John Phillips: Agreed as important and requiring changes.

Addressed by new content addressing verifiable history. See Merge !68

Issue #104: DFGT S-05 - Definition of Delegated National Register Operator Role​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/104

Original Comment​

Document​

Glossary (Operator) / Implementation Guidelines Para 2.6

Issue​

The Glossary defines 'Operator' as the UNICC-appointed technical operator of GRID services. It does not separately define the entity that operates a national register under government delegation. In certain jurisdictions, the legal mandate-holder and the day-to-day operator of the register may be different bodies - for example, a ministry holding the legal mandate and a state agency or commercial entity operating the digital system. The GRID onboarding pathway at Para 2.6 does not distinguish these roles. A GRID record that names only one body conflates legal accountability with operational responsibility.

Proposal​

Define a separate role - 'National Register Operator' or 'Delegated Register Operator' - for a body operating a register under formal authority delegated from the Authoritative Registrar. The GRID metadata schema should accommodate both roles, with the Authoritative Registrar designated as the legally accountable party and the Operator designated for operational and technical contact. 'Operator' in the Glossary should be reserved for the UNICC role only, with a separate defined term for the national delegation context.

Discussion​

[2026-10-02] John Phillips: The observation is understandable in that, for many jurisdictions, there may be a separation of legal authority and operating responsibility.

However as we articulate in Article 3 "Attribution and Liability"

"For data marked as Authoritative, the Registrar is legally responsible for the accuracy of the data that they natively issue and control."

This answers the question. Rather than being interested in the difference between "legal accountability and operational responsibility", we are solely interested in legal accountability.

The legal delegation of operating responsibility within the UN Member State is rightly at their discretion. The GRID needs to know who is legally accountable.

However, in order to avoid potential confusion, the term "Operator" in the Glossary will be prefixed by "GRID" and hence become "GRID Operator".

[2026-10-02] John Phillips: This change will be made in PRU04 and brought into main when the PRU04 branch is merged.

Issue #103: DFGT S-04 - Objective Criteria for Authoritative Registers​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/103

Original Comment​

Document​

Glossary (Authoritative Register) / Implementation Guidelines Para 2.5

Issue​

The Glossary defines 'Authoritative Register' as a register that serves as a trusted and authoritative source. This is circular. Implementation Guidelines Para 2.5 preserves a distinction between register, Registrar, registry data and GRID record but does not state what makes a register authoritative. Para 2.2 explicitly disclaims superiority between register types. There are no objective criteria by which a genuinely disputed case - two government bodies both claiming to be the official company registry - could be resolved under the current definitions.

Proposal​

Replace the Glossary definition with criteria-based language. An Authoritative Register should be defined by reference to: (i) a statutory or legal mandate conferring competence over the register type in question; (ii) legal effect of entries in the register under the applicable law; (iii) designated accountability and governance for accuracy, currency and status of the data; and (iv) identifiable public accountability mechanisms. The eligibility verification step at Para 2.6 should reference these criteria as the basis for admissions decisions.

Discussion​

[2026-10-01] Sankarshan Mukhopadhyay: I agree that the current definition appears to be circular, but I would be cautious about turning GRID's definition into a universal test by which competing claims of public authority are resolved.

The stronger model may be to define an Authoritative Register as one whose authority is traceable to an identifiable legal or institutional basis within the relevant jurisdiction and scope. The GRID record can then expose the evidence needed to understand that basis, including the responsible authority, legal instrument or mandate, register function, jurisdiction, scope, and applicable period.

The proposed criteria around mandate, legal effect, accountability, accuracy, and currency are useful admission evidence. They should not imply that GRID itself adjudicates a constitutional, statutory, or inter-agency dispute about which public body is authoritative where the applicable jurisdiction has not resolved that question.

In other words, GRID can make the basis and scope of claimed authority inspectable and verifiable without becoming the source of that authority.

[2026-10-02] John Phillips: The existing glossary definition is:

Authoritative Register
A register that, under the law of a jurisdiction, serves as a trusted and authoritative source of official information. See also Register.

This will be changed to

Authoritative Register
A register that, under the law of a UN Member State jurisdiction, serves as the legal source of official register information and whose authority is traceable to an identifiable legal or institutional basis within the relevant jurisdiction and scope. See also Register.

These changes have been incorporate in the PRU04 branch and will be merged with main in due course.

Issue #102: DFGT S-03 Harmonization of Data Quality Tiers​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/102

Original Comment​

Document​

DIA Legal Meaning and Data Structure Para 2.3 vs Implementation Guidelines Para 3.8

Issue​

The DIA document Para 2.3 defines three data quality tiers: Authoritative, Verified, and Asserted. The Implementation Guidelines Para 3.8 defines two: Verified and Asserted, with no mention of Authoritative. These two provisions are in conflict. A verifier cannot know from the Implementation Guidelines whether Registrar-controlled content is in a separate category or subsumed under Verified. Para 3.8 states that this distinction is 'necessary to preserve legal certainty and to avoid unintended extension of Registrar responsibility' - yet the standard itself is inconsistent on this point.

Proposal​

Harmonise to a single model across all documents. The three-tier model (Authoritative / Verified / Asserted) is the more precise of the two and should be adopted consistently. Each tier should be defined with a machine-readable representation in the schema, not described only in prose. Verifier guidance should be added stating the legal implications of each tier for reliance purposes.

Discussion​

[2026-10-02] John Phillips: Agreed.

Changes made to the implementation guidelines in branch PRU04 and will be merged with main.

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/101

Original Comment​

Document​

DIA Legal Meaning and Data Structure Paras 1.2 and 2.1 / Implementation Guidelines Para 1.3

Issue​

No normative statement in any document explicitly distinguishes between (i) verification of legal identity or registration status, (ii) verification of control of a cryptographic identifier, and (iii) verification of authority to act on behalf of a legal person. Implementation Guidelines Para 1.3 addresses identifier equivalence but does not address the specific risk of key control being treated as a proxy for identity. This gap is material because keys may be provisioned by a Registrar, borrowed, stolen or shared, and because personnel checking credentials in trade contexts are not typically legal experts.

Proposal​

May add the following explicit statement to Para 1.3 of the Implementation Guidelines and to Section 3.1 of the DIA document: 'Control of a cryptographic identifier shall not, by itself, be interpreted as evidence of legal identity, registration status or authority to represent the subject entity. Verification of key control is a technical assurance only. Legal identity and authority to act on behalf of a legal person are separate determinations governed by the applicable law and the underlying register.'

Discussion​

[2026-10-01] Sankarshan Mukhopadhyay: I support the proposed distinction and would make it slightly more general.

Proof of control of a cryptographic identifier establishes a technical relationship between an actor and a verification method. It does not, without additional evidence, establish who that actor is in law, whether the represented entity currently has the claimed registration status, or whether the actor is authorised to exercise a particular power on behalf of that entity.

I would therefore keep three propositions explicitly separate throughout the model:

  1. identity or registration: who or what the authoritative register recognises;
  2. cryptographic control: who can demonstrate control of the relevant verification method; and
  3. authority: what the controlling actor is permitted to do, for whom, within what scope, and during what period.

A verifier may need evidence for all three, but success in one should not be treated as proof of the others. This also frames the delegation discussion in #108 with a cleaner architectural foundation.

[2026-10-02] John Phillips: Agree with this observations, but I sense that an element of "solution in mind" has been injected in the proposed text. In addition, these are considerations that go to the Registry's own legal and operating processes, over which GRID has no control.

My consideration is that it is possible to imagine a system where control of the cryptographic identifier is sufficient proof of identity, where (for example) biometrically bound credentials and local, secure, biometric processes are part of the control process. But the real answer is that "it depends" - as in it depends on the jurisdiction's own definition of what constitutes sufficient proof of identity.

So, while the comment rightly states that control cannot be assumed to mean proof of legal identity, not can it be assumed NOT to mean proof of legal identity. It depends on the solution AND legal environment.

We will adopt the proposed text with changes. Rather than:

"Control of a cryptographic identifier shall not, by itself, be interpreted as evidence of legal identity, registration status or authority to represent the subject entity. Verification of key control is a technical assurance only. Legal identity and authority to act on behalf of a legal person are separate determinations governed by the applicable law and the underlying register."

we will add:

"Note that control of a cryptographic identifier MAY not, by itself, provide sufficient evidence of legal identity, registration status or authority to represent the subject entity within a UN Member State jurisdiction. Legal identity and authority to act on behalf of a legal person are determinations governed by the UN Member State applicable law and the underlying register."

This will be added to the DIA (Introduction) and Implementation Guidelines documents in branch PRU04 before being merged into main.

Issue #100: DFGT S-01 - Distinct Naming for Credential Levels​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/100

Original Comment​

Document​

Pilot Design Document Para 1 / DIA Legal Meaning and Data Structure Para 1.2

Issue​

Two distinct credential types are both called a DIA. The Pilot Design Document states that GRID issues a DIA for a Registrar. The DIA Legal and Data Structure document states that a Registrar issues a DIA for its registrant. These are different signing authorities, different subjects, different purposes, and different liability positions, yet both outputs carry the same name. A verifier examining a credential cannot determine from its label which type it is or who bears responsibility for its content.

Proposal​

Introduce distinct names for the two credential levels. Suggested: Registrar Trust Credential (RTC) for the GRID-to-Registrar attestation; Digital Identity Anchor (DIA) reserved for the Registrar-to-entity credential. Apply these names consistently across all documents and the schema. Where a trust hierarchy is described, label each level explicitly.

Discussion​

[2026-10-09] John Phillips: Thank you for this review and for highlighting the need for clarity in the trust hierarchy. You are absolutely correct that the relationship and liability models between GRID/Registrar and Registrar/Entity are distinct and must be easily distinguishable to the reader.

However, we will not be adopting the proposal to create a separate "Registrar Trust Credential (RTC)" type, due to how this fits into the broader UN Transparency Protocol (UNTP) and Verifiable Credential architecture.

Technical Rationale: The Digital Identity Anchor (DIA) is a standardized Verifiable Credential type defined by UNTP. In the Verifiable Credentials data model, the specific role, liability, and context of a credential are not determined by its top-level semantic label (the type), but rather by the contents of the payload—specifically the issuer and credentialSubject fields—verified against a Trust Registry.

Introducing new credential types based on the role of the issuer (e.g., an RTC for Registrars, a hypothetical "Sub-RTC" for sub-registrars, etc.) would lead to an unnecessary explosion of credential types and complicate the technical maintenance of the schemas. Software verifiers already know exactly who bears responsibility for the credential by reading the issuer DID; they do not need a different credential name to process the trust chain.

Proposed Resolution: While the technical schema will remain as a single DigitalIdentityAnchor type, we agree that the documentation needs to be clearer for human readers.

To resolve this issue, we will update the prose across the Pilot Design Document and the DIA Legal Meaning documents. We will establish a naming convention in the text using qualifying adjectives—such as "GRID-issued DIA" (or Registrar DIA) and "Registrar-issued DIA" (or Entity DIA)—whenever the trust hierarchy is being described.

Issue #99: DM Comment 8 - Clarify the limits of reliance on directory inclusion​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/99

Original Comment​

Comment 8 Clarify the limits of reliance on directory inclusion​

Location. Implementation Guidelines 1.1, 1.4, 1.8, 2.5, 3.1 and 3.6.3.

Page URL

Issue. The Guidelines already preserve registrar authority and distinguish official information from other assertions. A short statement would help downstream product systems apply these boundaries.

Proposed wording. Inclusion in GRID helps users discover a participating registrar and its register, and provides information relevant to trust. Inclusion alone does not certify every claim associated with a registered entity or identifier, establish product conformity, or prove the physical authenticity of an item bearing that identifier.

The legal or evidentiary effect of a registration, extract, attestation or Digital Identity Anchor continues to derive from the underlying register, the issuer’s competence and applicable law.

Rationale. This preserves the significance of official register outputs while helping users check whether a particular attestation supports the specific product claim and decision.

Discussion​

[2026-10-09] John Phillips: Thank you for this comment. We agree that this is a necessary and valuable clarification.

While the Legal Governance and Target Operating Model clearly establishes that the GRID acts solely as a discovery layer for Authoritative Registrars, and the Implementation Guidelines (Principles 1.1, 1.4, 1.8) establish that GRID does not alter the underlying legal effect of the Register, these documents currently focus primarily on the relationship between the GRID and the Registrars.

We agree that an explicit boundary needs to be explained for downstream relying parties. Because DIAs and GRID verifications will be used in supply chains alongside product-level credentials (e.g., Digital Product Passports, Certificates of Conformity), there is a risk of a "halo effect" where a verified entity identity is mistakenly interpreted as an endorsement of that entity's physical products or claims.

We will adopt your proposed wording to ensure this limitation is explicitly clear to implementers and relying parties.

Action: We will integrate the following text into the Implementation Guidelines (likely as a new common principle or as a distinct subsection regarding "Limits of Reliance"):

"Inclusion in GRID helps users discover a participating registrar and its register, and provides information relevant to trust. Inclusion alone does not certify every claim associated with a registered entity or identifier, establish product conformity, or prove the physical authenticity of an item bearing that identifier. The legal or evidentiary effect of a registration, extract, attestation, or Digital Identity Anchor continues to derive from the underlying register, the issuer’s competence, and applicable law."

Thank you for helping to strengthen the legal clarity and practical implementation of the GRID framework.

Issue #98: DM Comment 7 - Specify the material needed to verify signed outputs​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/98

Original Comment​

Comment 7 Specify the material needed to verify signed outputs​

Location. GRID Data Schema, DID / Public Key field; Implementation Guidelines 2.3, 2.6, 3.7 and 3.9.

Page URLs

Issue. The guidance addresses key management, and the schema includes did. Implementers would benefit from explicit rules linking each signed output to its authorised verification method.

Proposed wording. Distinguish the registrar’s identifier from its verification-method references, and state the purposes for which each key is authorised. Cover both signed registrar metadata and DIA issuance where applicable.

Define how verification material is retrieved or packaged, how rotation and compromise are reported, when changes take effect and how authenticated historical references are retained or located.

State the verification-material requirements for each relevant profile, while preserving baseline listing for eligible registrars that do not issue credentials.

Rationale. A profile requiring signed outputs needs a clear verification path. Historical checks require authenticated evidence linking a key to its issuer at the relevant time. An issuer’s stated issuance date alone does not establish that a signature predates a key compromise.

Discussion​

[2026-10-08] John Phillips: Thank you for this critical feedback. We have tagged as the related to history - along with a number of other issues on this topic.

We agree that coupling a registrar's GRID participation status with their legal mandate and verification-key statuses creates risks for historical reliance and non-repudiation.

A voluntary withdrawal from the GRID should end active discovery but must not implicitly invalidate previously issued DIAs or obscure the historical validity of a registrar’s mandate and keys.

With specific reference to the concerns you raise, the following updates are proposed:

Implementation Guidelines (2.6 & 2.7): We will update the onboarding and technical implementation pathways to explicitly decouple these three lifecycles (Participation, Mandate, and Key status). We will also clarify the handling of temporal boundaries (e.g., mandates with no fixed end date vs. unknown end dates, and suspended vs. expired mandates).

Governance Appendix A.4: We will define the retention policies for historical metadata. This will clarify how the GRID references or retains historical keys and mandate records for withdrawn participants solely to support the review of past decisions, ensuring they are clearly marked as inactive/withdrawn for current discovery.

GRID Data Schema: Where necessary, we will ensure the schema supports these decoupled status flags and temporal validity periods.

We need to consider these changes in parallel with the other changes that have been recommended and/or become self-evidently needed to address historical data verification.

[2026-10-15] John Phillips: Addressed by new content addressing verifiable history. See Merge !68

Issue #97: DM Comment 6 - Align participation profiles across the documents​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/97

Original Comment​

Comment 6 Align participation profiles across the documents​

Location. GRID Data Schema, Participant Profile field; Implementation Guidelines 3.10 and 4.1 to 4.5; Governance 5.

Page URLs

Issue.

The schema accepts A to E for participant_profile. The Guidelines define E as issuance outside GRID participation, while Governance 5 illustrates levels A to C. These uses need an explicit relationship.

Proposed wording. Define one consistent set of profile codes, or publish an explicit mapping between maturity levels and implementation profiles. Distinguish GRID participation from technical capability.

Clarify whether Profile E may occur in a published GRID record and, if so, for what purpose. It must not imply admission to GRID. If the schema is intended only for admitted participants, exclude Profile E from that use or provide a separate representation for non-participating issuers.

Specify whether each DIA-capable profile also requires signed and harvested registrar metadata, and align those requirements across the schema and guidance.

Rationale. This makes profile codes usable by software without confusing technical conformity, digital maturity and admission to GRID.

Discussion​

[2026-10-08] John Phillips: Closed with #111

Issue #96: DM Comment 5 - Correct the country field in the published example​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/96

Original Comment​

Comment 5 Correct the country field in the published example​

Location. GRID Data Schema, Formal Definition and Example.

Page URL

Issue. The published schema requires UN_member_state, but the example supplies un_member_state. JSON property names are case-sensitive, so the example lacks the required property.

Proposed wording. Change the example’s country property from un_member_state to UN_member_state. Check the complete example against registrar-v1.json, and keep published examples consistent with the schema when either is updated.

Rationale. This removes a directly reproducible mismatch that implementers could otherwise copy into their own records.

Discussion​

[2026-09-22] John Phillips: This is a typo - we should just fix this in the next merge.

[2026-09-22] John Phillips: Made the change in the https://opensource.unicc.org/un/unece/uncefact/gtr/-/tree/PublicReviewCommentsUpdate02?ref_type=heads branch

Issue #95: DM Comment 4 - Clarify subdivision codes and jurisdictional scope​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::1-discussion URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/95

Original Comment​

Comment 4 Clarify subdivision codes and jurisdictional scope​

Location. GRID Data Schema, Subdivision Code field and Formal Definition.

Page URL

Issue. The subdivision_code pattern accepts a suffix such as BC, while the description refers to ISO 3166-2. The schema permits omission or null, while the table directs use of null when no subdivision exists.

Proposed wording. State whether subdivision_code holds the subdivision suffix, read together with UN_member_state, or a full ISO 3166-2 code. Align the description, validation rule and examples with the chosen format.

Define whether omission and null have the same meaning. Distinguish the territory in which a registrar has authority from its office address, and document how to reference multiple or overlapping scopes where relevant.

Where cross-border recognition is relevant, allow a reference to the recognition instrument, including its territorial and functional scope and applicable period. Recognition should not imply that a domestic mandate automatically transfers to another jurisdiction.

Rationale. These clarifications help systems interpret an existing field consistently and match authority to the relevant jurisdiction and registration purpose.

Discussion​

[2026-10-08] John Phillips: Thank you for your review and for raising these excellent points. The project team agrees with your assessment; there are conflicting definitions in the current draft documentation and schema.

To resolve these ambiguities, we will update the schema (registrar.json) and the Qualities of Registrar Data document as follows:

  1. Format Alignment: We will align on the full ISO 3166-2 format (e.g., XX-YYY). We will update the regex pattern to ^[A-Z]{2}-[A-Z0-9]{1,3}$, and update the text/examples to reflect the full string (e.g., CA-BC), removing the need to read it in conjunction with UN_member_state.
  2. Null vs. Omission: We will clarify the semantics by adding the subdivision field to the required array so explicit null is mandatory.
  3. Jurisdiction vs. Address: We agree that jurisdictional scope of authority may be distinct from physical or administrative office location. After review, no change was made. While we understand the observation, the physical address of the registrar is technically speaking irrelevant. UN Member State legislation defines the Registrar's authority and geographic scope. If the same registrar operates over several legally independent states, and each wishes to participate in the GRID, then several entries can be made in the GRID for the same registrar by different UN Member States.
  4. Cross-border/Complex Scopes: See 3.

We will push these schema changes shortly. Thank you for helping improve the GRID specification.

Status​

State: closed | Author: John Phillips Labels: public-review-comment::1-discussion URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/94

Original Comment​

Location. GRID Data Schema, Legal Basis and Authoritative Body fields; Implementation Guidelines 2.5 and 3.9.

Page URLs

Issue. A legal-basis or oversight reference may not explain whether another organisation exercises delegated legal authority or simply provides technical services.

Proposed wording. Allow an optional structured reference to an authoritative instrument or credential recording a delegation. Where available, identify the delegator, delegate, delegated function, register or claims covered, jurisdiction and applicable period.

Distinguish delegated legal authority from technical service provision. A missing delegation reference should not, by itself, imply either that a legally valid delegation exists or that it does not exist.

Rationale. This helps users locate the evidence needed to assess a delegation. It does not change admission criteria or make GRID responsible for deciding every subsequent authority relationship. Legal effect remains governed by the relevant instrument and applicable law.

Discussion​

[2026-10-08] John Phillips: Thank you for your comment and the proposal on how to handle delegated legal authority.

While we completely agree on the need to clearly distinguish between an entity exercising delegated legal authority and one that is simply providing technical services, the project team prefers to maintain a simpler, flatter approach to how this is represented in GRID.

Our primary goal is to identify exactly who is legally authorized to register within a UN Member State. We expect that specific party to be the signatory to the registrations.

Rather than creating a nested hierarchy of delegation relationships and using credentials to map out delegators and delegates, our approach is to focus on the end result of that delegation. If authority has been delegated, GRID will simply:

  1. Register the body to which the authority has been delegated as the authoritative body.

  2. Register the specific legislation, instrument, or legal basis that enacts this delegation.

Attempting to capture structured details of the full delegation chain (delegator, delegate, periods, jurisdictions, etc.) risks overcomplicating the schema. It would also require GRID to echo too much of the local, internal administrative structures of individual UN Member States, which falls outside our intended scope.

Therefore, we do not plan to add the proposed complex structured reference for delegations to the data schema. However, we take your point regarding the potential confusion between legal delegates and technical service providers. To address this, we can look at clarifying the Implementation Guidelines to explicitly state how delegations should be handled—specifically, by registering the delegated entity as the authoritative body alongside the legislation that grants them that power.

Issue #93: DM Comment 2 - Define authenticated directory versions and updates​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::1-discussion URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/93

Original Comment​

Comment 2 Define authenticated directory versions and updates​

Location. Implementation Guidelines 2.3 and 2.7; supporting architecture in About the GRID and DIAs 3.1.

Page URLs

Issue. The design already supports downloaded directory copies and avoids a query for every transaction. Shared rules for publishing and updating those copies would make this approach easier to implement consistently.

Proposed wording. Define how complete directory copies, or snapshots, are authenticated and versioned. Where a change feed is offered, specify how it works with those snapshots.

Specify a stable version identifier, publication time, integrity protection, publisher authentication, update order and a way to detect missing updates. Each change should identify the affected entry, its effective time and any record it replaces or withdraws. Retain enough withdrawal information to update cached copies correctly.

Explain how to recover after missed updates and prevent an older version from being passed off as current. Publish refresh expectations. Users should retain the version relied upon and apply explicit limits on its age and coverage.

Authentication establishes a snapshot’s source and integrity; it does not establish that no newer version exists. Baseline participants should be supported through the directory operator’s authenticated publication process.

Rationale. This supports reliable local verification without requiring GRID contact for every transaction. The detailed publication rules could sit in a technical profile, with the functional requirement retained in the Guidelines.

Discussion​

[2026-10-02] John Phillips: We agree that Relying Parties need a way to authenticate full directory snapshots for offline verification without needing to query the GRID Hub for every transaction.

Because our current pilot design is built on a Git-backed CI/CD architecture, the source of truth is versioned and signed natively. To pass this assurance down to the downloaded snapshots (like the CSV), we propose extending our current cryptographic pipeline rather than building a new query system.

Proposed Solution (Aligning with current Pilot Architecture):​
  1. Signed Snapshots: We extend the existing generate-grid.js script. After it generates the W3C Verifiable Credentials (DIAs) and the aggregate CSV directory, it will package a snapshot-manifest.json.

  2. Manifest Contents: This manifest will contain the Publication Time, a stable Version Identifier (derived from the Git commit hash / Pipeline ID), and the cryptographic hash of the CSV directory.

  3. Authentication: The manifest will be cryptographically signed by the GRID Hub's private key (using the same Linked Data proofs we use for the DIAs).

  4. Rollback & Update Management: By including sequential version IDs and timestamps in the signed manifest, Relying Parties can configure their local systems to reject older versions (preventing rollback attacks) and establish explicit limits on cache age. Withdrawals or changes naturally propagate on the next CI/CD run and will be reflected in the newest signed snapshot.

This "sign everything" approach fulfills the requirements of Issue 93 locally, while keeping the GRID Hub architecture lightweight and decentralized.

Issue #92: DM Comment 1 - Separate participation status from mandate and key status​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/92

Original Comment​

Comment 1 Separate participation status from mandate and key status​

Location. Implementation Guidelines 2.6 and 2.7; Governance Appendix A.4.

Page URLs

Issue. Withdrawal ends active discovery through GRID. Users also need to distinguish changes in a registrar’s legal mandate and verification keys. These events can have different effects on present and past reliance.

Proposed wording. GRID data and guidance should distinguish participation status, the status and scope of the registrar’s legal mandate, and verification-key status. Where GRID does not hold the authoritative record for a mandate or key, it may reference the responsible source.

Identify that source, when a status takes effect and when it was published. Distinguish a mandate with no fixed end date from one whose end date is unknown, and a suspended mandate from an expired one.

Define what historical information may be retained or referenced under the governance arrangements. Such records should support review of past decisions without presenting a withdrawn registrar as an active participant or as currently verified through GRID.

Rationale. This allows an earlier reliance decision to be reconstructed while preserving withdrawal rights. Current decisions must still consider relevant later findings and status information.

Discussion​

[2026-10-01] Sankarshan Mukhopadhyay: I agree with separating participation status, mandate status, and verification-key status, but I think the temporal model needs one further distinction.

For historical use, it may be useful to distinguish:

  1. what GRID or another authoritative source recorded at the historical time;
  2. what the relevant mandate, key, or participation state is subsequently understood to have been effective at that time; and
  3. the determination a verifier makes now, using the evidence and rules available now, about that historical state.

Those propositions can diverge. A compromise, withdrawal, suspension, or other event may be discovered and recorded after the date from which it is considered effective.

I would therefore avoid defining historical support only as retention of earlier records or snapshots. Those records are useful evidence, but they do not necessarily represent the final effective state for the period being examined.

The guidance could instead require enough temporal information to distinguish recorded time, effective time, and evaluation time, while leaving the underlying persistence mechanism to implementations.

[2026-10-13] John Phillips: Agreed that we need to make changes to address this issue.

The consideration is :

  1. A Registrar's legal mandate is determined by domestic statute
  2. its cryptographic keys are controlled by technical custodians, and
  3. its participation status is governed by the GRID Board

These are three distinct types of "status". Lumping them into a single status property leads to errors (e.g., treating a withdrawn Registrar as legally illegitimate, or assuming a valid mandate implies an uncompromised key).

Propose to adopt this separation across all GRID schemas and governance documents.

Changes will be made in the PRU04 branch before proposing a merge.

Successful review and merge will close this issue.

Issue #91: David Male - RAMIFY Public Review - Set of Comments​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/91

Original Comment​

These are the Public Review Comments received from David Make of RAMIFY OS. The review comments were received as a PDF attachment sent to the GRID Project Email and received on 21 Sept. The email link is: https://groups.io/g/UNCEFACTGlobalTrustRegistry/topic/gtr_public_review_feedback/121366028

The contents of the email and document have been replicated below for transparency.

There are a number of comments in the document. These will be separated into separate comments and managed as sub-tasks (child-tasks) of this parent issue.


Email​

Re: GRID

See 4-page PDF submission attached.

This includes the following details for each comment: the page URL, the issue, our proposed solution, and our rationale.

In the meantime, I'll also sign up for an account through: https://opensource.unicc.org/users/sign_up

UNICC OpenSource Code Repositories

Warm Regards

David Male Founder, RAMIFY OS

Comments​

GRID public review submission​

Global Registry Information Directory

David Male | Founder, RAMIFY OS | Hobart, Australia

21 September 2026 | Version 1.5 | Public review submission

Purpose and overall position​

RAMIFY OS is an Australian venture developing software to assess physical-product claims using evidence, externally established authority and defined policies. Our intended users include procurement platforms, marketplaces and AI systems.

We support GRID’s role in helping users discover authoritative registrars while preserving the authority of the underlying registers. The comments below propose targeted clarifications for consistent interpretation, reliable local use and reproducible decisions. They support phased implementation and proportionate requirements for participants.

The comments refer to the Stage 4 public-review web edition checked on 21 September 2026. Section references are to the web pages identified under each comment. Proposed wording is offered for the project team’s consideration.

Comment 1 Separate participation status from mandate and key status​

Location. Implementation Guidelines 2.6 and 2.7; Governance Appendix A.4.

Page URLs

Issue. Withdrawal ends active discovery through GRID. Users also need to distinguish changes in a registrar’s legal mandate and verification keys. These events can have different effects on present and past reliance.

Proposed wording. GRID data and guidance should distinguish participation status, the status and scope of the registrar’s legal mandate, and verification-key status. Where GRID does not hold the authoritative record for a mandate or key, it may reference the responsible source.

Identify that source, when a status takes effect and when it was published. Distinguish a mandate with no fixed end date from one whose end date is unknown, and a suspended mandate from an expired one.

Define what historical information may be retained or referenced under the governance arrangements. Such records should support review of past decisions without presenting a withdrawn registrar as an active participant or as currently verified through GRID.

Rationale. This allows an earlier reliance decision to be reconstructed while preserving withdrawal rights. Current decisions must still consider relevant later findings and status information.

Comment 2 Define authenticated directory versions and updates​

Location. Implementation Guidelines 2.3 and 2.7; supporting architecture in About the GRID and DIAs 3.1.

Page URLs

Issue. The design already supports downloaded directory copies and avoids a query for every transaction. Shared rules for publishing and updating those copies would make this approach easier to implement consistently.

Proposed wording. Define how complete directory copies, or snapshots, are authenticated and versioned. Where a change feed is offered, specify how it works with those snapshots.

Specify a stable version identifier, publication time, integrity protection, publisher authentication, update order and a way to detect missing updates. Each change should identify the affected entry, its effective time and any record it replaces or withdraws. Retain enough withdrawal information to update cached copies correctly.

Explain how to recover after missed updates and prevent an older version from being passed off as current. Publish refresh expectations. Users should retain the version relied upon and apply explicit limits on its age and coverage.

Authentication establishes a snapshot’s source and integrity; it does not establish that no newer version exists. Baseline participants should be supported through the directory operator’s authenticated publication process.

Rationale. This supports reliable local verification without requiring GRID contact for every transaction. The detailed publication rules could sit in a technical profile, with the functional requirement retained in the Guidelines.

Location. GRID Data Schema, Legal Basis and Authoritative Body fields; Implementation Guidelines 2.5 and 3.9.

Page URLs

Issue. A legal-basis or oversight reference may not explain whether another organisation exercises delegated legal authority or simply provides technical services.

Proposed wording. Allow an optional structured reference to an authoritative instrument or credential recording a delegation. Where available, identify the delegator, delegate, delegated function, register or claims covered, jurisdiction and applicable period.

Distinguish delegated legal authority from technical service provision. A missing delegation reference should not, by itself, imply either that a legally valid delegation exists or that it does not exist.

Rationale. This helps users locate the evidence needed to assess a delegation. It does not change admission criteria or make GRID responsible for deciding every subsequent authority relationship. Legal effect remains governed by the relevant instrument and applicable law.

Comment 4 Clarify subdivision codes and jurisdictional scope​

Location. GRID Data Schema, Subdivision Code field and Formal Definition.

Page URL

Issue. The subdivision_code pattern accepts a suffix such as BC, while the description refers to ISO 3166-2. The schema permits omission or null, while the table directs use of null when no subdivision exists.

Proposed wording. State whether subdivision_code holds the subdivision suffix, read together with UN_member_state, or a full ISO 3166-2 code. Align the description, validation rule and examples with the chosen format.

Define whether omission and null have the same meaning. Distinguish the territory in which a registrar has authority from its office address, and document how to reference multiple or overlapping scopes where relevant.

Where cross-border recognition is relevant, allow a reference to the recognition instrument, including its territorial and functional scope and applicable period. Recognition should not imply that a domestic mandate automatically transfers to another jurisdiction.

Rationale. These clarifications help systems interpret an existing field consistently and match authority to the relevant jurisdiction and registration purpose.

Comment 5 Correct the country field in the published example​

Location. GRID Data Schema, Formal Definition and Example.

Page URL

Issue. The published schema requires UN_member_state, but the example supplies un_member_state. JSON property names are case-sensitive, so the example lacks the required property.

Proposed wording. Change the example’s country property from un_member_state to UN_member_state. Check the complete example against registrar-v1.json, and keep published examples consistent with the schema when either is updated.

Rationale. This removes a directly reproducible mismatch that implementers could otherwise copy into their own records.

Comment 6 Align participation profiles across the documents​

Location. GRID Data Schema, Participant Profile field; Implementation Guidelines 3.10 and 4.1 to 4.5; Governance 5.

Page URLs

Issue.

The schema accepts A to E for participant_profile. The Guidelines define E as issuance outside GRID participation, while Governance 5 illustrates levels A to C. These uses need an explicit relationship.

Proposed wording. Define one consistent set of profile codes, or publish an explicit mapping between maturity levels and implementation profiles. Distinguish GRID participation from technical capability.

Clarify whether Profile E may occur in a published GRID record and, if so, for what purpose. It must not imply admission to GRID. If the schema is intended only for admitted participants, exclude Profile E from that use or provide a separate representation for non-participating issuers.

Specify whether each DIA-capable profile also requires signed and harvested registrar metadata, and align those requirements across the schema and guidance.

Rationale. This makes profile codes usable by software without confusing technical conformity, digital maturity and admission to GRID.

Comment 7 Specify the material needed to verify signed outputs​

Location. GRID Data Schema, DID / Public Key field; Implementation Guidelines 2.3, 2.6, 3.7 and 3.9.

Page URLs

Issue. The guidance addresses key management, and the schema includes did. Implementers would benefit from explicit rules linking each signed output to its authorised verification method.

Proposed wording. Distinguish the registrar’s identifier from its verification-method references, and state the purposes for which each key is authorised. Cover both signed registrar metadata and DIA issuance where applicable.

Define how verification material is retrieved or packaged, how rotation and compromise are reported, when changes take effect and how authenticated historical references are retained or located.

State the verification-material requirements for each relevant profile, while preserving baseline listing for eligible registrars that do not issue credentials.

Rationale. A profile requiring signed outputs needs a clear verification path. Historical checks require authenticated evidence linking a key to its issuer at the relevant time. An issuer’s stated issuance date alone does not establish that a signature predates a key compromise.

Comment 8 Clarify the limits of reliance on directory inclusion​

Location. Implementation Guidelines 1.1, 1.4, 1.8, 2.5, 3.1 and 3.6.3.

Page URL

Issue. The Guidelines already preserve registrar authority and distinguish official information from other assertions. A short statement would help downstream product systems apply these boundaries.

**Proposed wording. ** Inclusion in GRID helps users discover a participating registrar and its register, and provides information relevant to trust. Inclusion alone does not certify every claim associated with a registered entity or identifier, establish product conformity, or prove the physical authenticity of an item bearing that identifier.

The legal or evidentiary effect of a registration, extract, attestation or Digital Identity Anchor continues to derive from the underlying register, the issuer’s competence and applicable law.

Rationale. This preserves the significance of official register outputs while helping users check whether a particular attestation supports the specific product claim and decision.

Discussion​

[No user comments on this issue yet]

Issue #90: Trade Finance DGFT - Set of Comments​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/90

Original Comment​

These are the Public Review Comments received from the Indian DGFT (Directorate General of Foreign Trade). The review comments were received as a Word Document attachment sent to the GRID Project Email and received on 21 Sept. The email link is: https://groups.io/g/UNCEFACTGlobalTrustRegistry/topic/gtr_public_review_feedback/121366028

The contents of the email and document have been replicated below for transparency.

There are a number of comments in the document. These will be separated into separate comments and managed as sub-tasks (child-tasks) of this parent issue.


Submitted byDirectorate General of Foreign Trade, Ministry of Commerce and Industry, Government of India
ContextIndia is implementing the BharatTradeNet (BTN) programme, a national digital trade infrastructure built on UNTP-compatible Verifiable Credentials and DIA architecture. These comments are offered from active implementation experience and are intended to strengthen the GRID standard before finalisation.
DateSeptember 21, 2026

The following submissions are offered in response to the Stage 4 Public Review. Each identifies a specific gap or inconsistency in the current documents, proposes a concrete resolution, and provides rationale grounded in the document text.

RefDocumentIssueProposal
S-01Pilot Design Document Para 1 / DIA Legal Meaning and Data Structure Para 1.2Two distinct credential types are both called a DIA. The Pilot Design Document states that GRID issues a DIA for a Registrar. The DIA Legal and Data Structure document states that a Registrar issues a DIA for its registrant. These are different signing authorities, different subjects, different purposes, and different liability positions, yet both outputs carry the same name. A verifier examining a credential cannot determine from its label which type it is or who bears responsibility for its content.Introduce distinct names for the two credential levels. Suggested: Registrar Trust Credential (RTC) for the GRID-to-Registrar attestation; Digital Identity Anchor (DIA) reserved for the Registrar-to-entity credential. Apply these names consistently across all documents and the schema. Where a trust hierarchy is described, label each level explicitly.
S-02DIA Legal Meaning and Data Structure Paras 1.2 and 2.1 / Implementation Guidelines Para 1.3No normative statement in any document explicitly distinguishes between (i) verification of legal identity or registration status, (ii) verification of control of a cryptographic identifier, and (iii) verification of authority to act on behalf of a legal person. Implementation Guidelines Para 1.3 addresses identifier equivalence but does not address the specific risk of key control being treated as a proxy for identity. This gap is material because keys may be provisioned by a Registrar, borrowed, stolen or shared, and because personnel checking credentials in trade contexts are not typically legal experts.May add the following explicit statement to Para 1.3 of the Implementation Guidelines and to Section 3.1 of the DIA document: 'Control of a cryptographic identifier shall not, by itself, be interpreted as evidence of legal identity, registration status or authority to represent the subject entity. Verification of key control is a technical assurance only. Legal identity and authority to act on behalf of a legal person are separate determinations governed by the applicable law and the underlying register.'
S-03DIA Legal Meaning and Data Structure Para 2.3 vs Implementation Guidelines Para 3.8The DIA document Para 2.3 defines three data quality tiers: Authoritative, Verified, and Asserted. The Implementation Guidelines Para 3.8 defines two: Verified and Asserted, with no mention of Authoritative. These two provisions are in conflict. A verifier cannot know from the Implementation Guidelines whether Registrar-controlled content is in a separate category or subsumed under Verified. Para 3.8 states that this distinction is 'necessary to preserve legal certainty and to avoid unintended extension of Registrar responsibility' - yet the standard itself is inconsistent on this point.Harmonise to a single model across all documents. The three-tier model (Authoritative / Verified / Asserted) is the more precise of the two and should be adopted consistently. Each tier should be defined with a machine-readable representation in the schema, not described only in prose. Verifier guidance should be added stating the legal implications of each tier for reliance purposes.
S-04Glossary (Authoritative Register) / Implementation Guidelines Para 2.5The Glossary defines 'Authoritative Register' as a register that serves as a trusted and authoritative source. This is circular. Implementation Guidelines Para 2.5 preserves a distinction between register, Registrar, registry data and GRID record but does not state what makes a register authoritative. Para 2.2 explicitly disclaims superiority between register types. There are no objective criteria by which a genuinely disputed case - two government bodies both claiming to be the official company registry - could be resolved under the current definitions.Replace the Glossary definition with criteria-based language. An Authoritative Register should be defined by reference to: (i) a statutory or legal mandate conferring competence over the register type in question; (ii) legal effect of entries in the register under the applicable law; (iii) designated accountability and governance for accuracy, currency and status of the data; and (iv) identifiable public accountability mechanisms. The eligibility verification step at Para 2.6 should reference these criteria as the basis for admissions decisions.
S-05Glossary (Operator) / Implementation Guidelines Para 2.6The Glossary defines 'Operator' as the UNICC-appointed technical operator of GRID services. It does not separately define the entity that operates a national register under government delegation. In certain jurisdictions, the legal mandate-holder and the day-to-day operator of the register may be different bodies - for example, a ministry holding the legal mandate and a state agency or commercial entity operating the digital system. The GRID onboarding pathway at Para 2.6 does not distinguish these roles. A GRID record that names only one body conflates legal accountability with operational responsibility.Define a separate role - 'National Register Operator' or 'Delegated Register Operator' - for a body operating a register under formal authority delegated from the Authoritative Registrar. The GRID metadata schema should accommodate both roles, with the Authoritative Registrar designated as the legally accountable party and the Operator designated for operational and technical contact. 'Operator' in the Glossary should be reserved for the UNICC role only, with a separate defined term for the national delegation context.
S-06Implementation Guidelines Paras 2.7 and 3.7Both Para 2.7 and Para 3.7 address current status, revocation, expiry and freshness only. Neither addresses the ability to determine whether a credential was valid at a specified past date. Trade disputes, financing decisions, customs audits, and proceedings before courts and arbitral tribunals require point-in-time verification: was this entity registered and in good standing on the date of shipment, or on the date the credit facility was drawn? A system that can only attest current status is architecturally incomplete for these purposes.Add to Para 3.7 (Technical Implementation Considerations for DIAs) a requirement to support temporal verification including: (i) issuer-signed timestamps enabling point-in-time proof of credential validity at the time of issuance; (ii) a status log mechanism or equivalent that preserves the state of a credential at defined historical points; (iii) verifier guidance on how historical status should be determined and presented for reliance purposes. Timestamping per RFC 3161 or an equivalent standard provides an established technical basis.
S-07Implementation Guidelines Para 3.7Para 3.7 lists 'credential lifecycle handling' and 'mechanisms for status, revocation, expiry or freshness' as capabilities DIA-capable implementation should support, but does not define what lifecycle events must be handled or how. The following events are unaddressed: (a) periodic key rotation; (b) key compromise, emergency revocation, and recovery; (c) entity restructuring - how an existing DIA relates to a successor entity following a merger or demerger; (d) entity termination - what status a DIA carries after the underlying registration lapses; (e) preservation of historical verifiability after identifier or key changes. Each Registrar will handle these independently if no model is provided, creating inconsistent outcomes across jurisdictions.Add a normative lifecycle model to Section 3.7 covering at minimum the five event types listed above, with required actions for the Registrar, required updates to the GRID record, required notification to affected credential holders, and required treatment of credentials already issued under the prior key or identifier. The model should address historical verifiability explicitly: a credential issued before a key change must remain verifiable through the issuer's publication of the prior verification material for a defined retention period.
S-08DIA Legal Meaning and Data Structure Para 1.2 / Implementation Guidelines Paras 1.6, 1.7 and Section 3.9The DIA document Para 1.2 describes an applicant-controlled DID model: the applicant provides the DID and the Registrar verifies control. Implementation Guidelines Paras 1.6 (Technological Neutrality) and 1.7 (Proportionality and Inclusive Maturity) imply that other provisioning models are permissible but do not state this. No implementation guidance addresses the registrar-provisioned model, in which the Registrar generates, manages and controls the cryptographic identifier on behalf of the registrant. Implementers reading the current documents may conclude that applicant-managed keys are required.Add an explicit statement to Section 3 that a registrar-provisioned cryptographic identifier is a permissible implementation of the DIA architecture, provided that a verifiable and trustworthy binding between the subject, the authoritative identifier and the cryptographic identifier is maintained. Extend Section 3.9 (Minimum Implementation Requirements) to include: the applicable control-verification process and its documentation for both the applicant-supplied and registrar-provisioned models. Reference Paras 1.6 and 1.7 as the governance basis for this flexibility.
S-09DIA Legal Meaning and Data Structure Para 2 / Implementation Guidelines generallyNeither the DIA document nor the Implementation Guidelines addresses how an authorised representative, officer, employee or agent of a registered entity is identified, credentialled and linked to that entity's DIA. In practice, a legal entity always acts through a human being with some scope of authority for some period. Without a standard model for this relationship, implementers will build incompatible local solutions, or rely on the informal assumption that whoever presents the entity DID is authorised to act - which repeats the conflation identified in S-02 above. A related privacy conflict: Article V of the DIA document prohibits personal information in the DIA, but delegation necessarily involves identifying a named person. These two requirements are in direct conflict without a resolution.Define a standard conceptual model for authority and delegation: (i) the entity-level DIA establishes the registered entity's identity; (ii) a separate, interoperable authority or delegation credential links a named individual or automated system to that entity, with defined scope, effective period, and revocation mechanism; (iii) the authority credential is verifiable independently of the entity DIA but resolves back to it. This need not be embedded within the DIA itself. The privacy conflict is resolved by this separation: the entity DIA contains no personal information, while the authority credential handles the delegation and can be subject to appropriate purpose limitation.
S-10Participating Registrar Data Schema (registrar-v1.json) / Implementation GuidelinesThe formal schema is already published as registrar-v1.json but no versioning or backward-compatibility strategy appears in any document. The schema will need to evolve as GRID participation grows, as new capabilities are added, and as requirements mature. Without a defined versioning governance framework, the first schema update will disrupt existing national integrations without warning. This risk increases with each new national participant that goes live under the current schema.Define and publish a versioning strategy covering: (i) version numbering convention (semantic versioning or equivalent); (ii) how new optional fields will be introduced without breaking existing parsers; (iii) how mandatory fields will be added with a defined minimum transition period; (iv) minimum deprecation notice before fields are removed; (v) how prior schema versions will remain resolvable for historical verification of records issued under those versions. This can be published as a short governance annex or as part of the GRID operating procedures.

Discussion​

[No user comments on this issue yet]

Issue #89: xiaoyu you - Trust level of registries and ratings​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/89

Original Comment​

This issue was raised by "xiaoyu you" on 15 Sept (AEST) via the online Google Form provided for public review comments. The details from the form have been transposed into this issue to enable processing by the project review team.


Page URL​

https://grid.unece.org/docs/draft/About/

Issue or Observation​

The trust level of different registries might be different. How to give a rating is a key issue.

Recommendation​

Maybe a rating mechanisms among users combined with the check for UN/CEFACT need to be considered.

Reviewer Name​

xiaoyu you

Reviewer Organisation​

CAICT

Reviewer Country​

CHINA

What email address can we use to contact you?​

provided by reviewer but not echoed here

Discussion​

[2026-09-22] John Phillips: This is an understandable request and one that has previously been discussed in the project development. Our rationale and decision is explained below.

It is clear that the methods used, the resources available, the rigour applied, and the legal framework amongst other factors will differ across different Registrars. It is also clear that this may result in other parties viewing the registrar as more or less trustworthy.

However, GRID made a decision early in the design process not to support a "rating" of Registrars as this was felt to be counter to the UN development goals. Our reasoning was:

  1. Each UN Member State has the right to decide which other Member States it finds trustworthy or not. The relative ranking that each UN Member State would apply to another Member State will differ depending on their relationship with the other Member State.

  2. Third parties already provide risk and credit scoring systems for many aspects of international trade. This can include Registrars. UN Member States and organisations operating in international trade may choose to take note of these rating systems.

  3. Therefore, the GRID will offer a simple, binary, process of eligibility assessment (as described in the operating model documentation). This process will realise a decision of eligible, or not.

No score will be applied.

Issue #88: JMFN Comment #5​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/88

Original Comment​

COMMENT 5 — DISPUTES AND CONFLICTING REGISTRATIONS​

Page URL: https://grid.unece.org/docs/draft/Documents/

Issue or observation:

Cross-border transactions may encounter disputed, duplicated or conflicting registrations, particularly concerning ownership, secured interests, intellectual property, legal representation or corporate authority.

Recommendation:

The implementation guidelines should define how GRID indicates that:

  1. A registration or registrar status is disputed.
  2. Multiple authorities claim overlapping competence.
  3. An appeal or administrative review is pending.
  4. A judicial or arbitral decision has modified the legal status.
  5. Users must consult the authoritative national system for final legal effect.

Rationale: A trustworthy global directory should communicate not only positive recognition but also uncertainty, conflict and the limits of reliance.

Discussion​

[2026-09-17] John Phillips: There are two levels of response here:

  1. Where it concerns membership of the GRID, the target operating model document describes the approach to dispute resolution.
  2. Where it concerns the local (within UN Member State) dispute, this is a matter for the UN Member State, not the GRID.

A GRID eligible Registry must be an authoritative UN Member State Registry nominated by the UN Member State's representative.

The GRID treats such nominations on face value - i.e. that they are the authoritative registry for the UN Member State in one or more areas - before checking that the nominated registry meets the eligibility requirements.

IMO this is covered by existing documentation.

Issue #87: JMFN Comment #4​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/87

Original Comment​

COMMENT 4 — ASSURANCE AND CERTIFIER CAPACITY​

Page URL: https://grid.unece.org/docs/draft/Documents/

Issue or observation:

Registrations and related credentials may be based on materially different assurance models, including self-declaration, administrative review, second-party assessment, independent inspection or accredited third-party certification.

Recommendation: Provide optional metadata or interoperability guidance indicating:

  1. The type of assurance.
  2. The evidence reviewed.
  3. The independence of the verifier.
  4. The verifier’s accreditation or legal authority.
  5. Whether physical inspection occurred.
  6. Whether the credential is subject to continuing monitoring.

Rationale: Making assurance levels visible would allow relying parties to distinguish an authoritative registration from an independently verified economic or physical fact.

Discussion​

[2026-09-17] John Phillips: The GRID requires UN Member State participating Registries to define the legal basis under which they operate. This is expected to be in the form of one or more links to the legislature of their UN Member State and is presented in the GRID along with other data elements describing the registry.

The registry's legal basis in their UN Member State would be where these questions are tackled, if they are tacked. These are not within scope of the GRID operation.

One of our key principles is that the GRID will not impact the way that a UN Member State Registrar registers items legally or commercially.

No change proposed to the GRID text.

Issue #86: JMFN Comment #3​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/86

Original Comment​

COMMENT 3 — DISTINGUISHING REGISTRATION FROM ECONOMIC EVIDENCE​

Page URL: https://grid.unece.org/docs/draft/About/

Issue or observation:

Verification that a registrar is authoritative does not, by itself, establish the current existence, value, availability, ownership, absence of encumbrances or physical custody of an underlying asset.

Recommendation:

Add a clear explanatory distinction among:

  1. Verification of the registrar’s authority.
  2. Verification of an identifier or registration.
  3. Verification of legal ownership or entitlement.
  4. Verification of physical existence or custody.
  5. Verification of valuation.
  6. Verification of commercial availability or deliverability.

External credentials should identify which of these propositions are actually being certified and which remain outside their scope.

Rationale:

This distinction would reduce the risk that users treat a valid digital credential as proof of facts that the issuing registrar did not verify.

Discussion​

[2026-09-17] John Phillips: I notice that this comment is directed at the "About" section which is not the normative (controlling specification) for the GRID. The points raised by Jose are sensible and relevant, and they are also addressed and/or defined as out of scope of the GRID's interests.

Taking the list as presented:

  1. Verification of the registrar’s authority.
    this is the core of the GRID's purpose and is tackled in the deliverable documentation, particularly the target operating model where eligibility requirements are described

  2. Verification of an identifier or registration.
    the GRID documentation tackles this in two ways: 1) by getting the GRID participant (the registrar) to describe what (if any) digital means exist to verify their issued identifiers and registration; 2) By encouraging the use of a verifiable credential (Digital Identity Anchor) both to provide a verifiable proof of GRID membership by the GRID, and as a format to be issued by the Registrar

  3. Verification of legal ownership or entitlement.
    Here I suspect there is more intent in Jose's question than is in the scope of the GRID. The GRID requires that participating Registrars provide the reference to the UN Member State legislation where the basis of their authority and the meaning of registration in their registry is defined. Importantly, we *do not expect all UN Member States to adopt the same legal framework. This also means that we do not prescribe a way in which the party registered by a GRID registrar can be determined to be the legal owner of the registration. Our scope of control is the GRID entries, not the entries of the Registrars listed in the GRID.

  4. Verification of physical existence or custody.
    Not in the scope of the GRID work. Some elements of UNTP may address this

  5. Verification of valuation.
    Not in the scope of the GRID work

  6. Verification of commercial availability or deliverability.
    Not in the scope of the GRID work. Some elements of UNTP and UNVTD may address this

Reading this comments, I wondered if we need text early in the "About" section that describes what the section is about and where to read the definitive (normative) content for the GRID. Reading the About page I see that this explanatory text already exists.

Issue #85: JMFN Comment #2​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/85

Original Comment​

COMMENT 2 — STATUS, REVOCATION AND HISTORICAL EVIDENCE​

Page URL: https://grid.unece.org/docs/draft/Documents/

Issue or observation: A registrar’s authority, jurisdiction, verification service or participation status can change. A verification that is valid today may no longer be valid when a transaction is audited or disputed in the future.

Recommendation: The GRID data model and implementation guidance should clearly support:

  1. Effective dates.
  2. Expiration dates.
  3. Suspension or revocation status.
  4. Reasons for status changes, where legally publishable.
  5. Version history.
  6. A method for proving what GRID reported at a particular historical date.

Rationale: Cross-border legal certainty requires more than current-status verification. Courts, auditors, financial institutions and supply-chain participants may need reliable historical evidence showing that a registrar was recognized when a credential or transaction was issued.

Discussion​

[2026-10-15] John Phillips: Agreed as important and requiring changes.

Addressed by new content addressing verifiable history. See Merge !68

Issue #84: JMFN Comment #1​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/work_items/84

Original Comment​

COMMENT 1 — INTEROPERABILITY WITH TRANSACTION AND ASSET CREDENTIALS​

Page URL: https://grid.unece.org/docs/draft/About/

Issue or observation: GRID correctly limits itself to identifying authoritative registrars and does not copy the contents of national registers. However, the documentation could explain more explicitly how a verified registrar identity may be referenced by external credentials representing assets, rights, invoices, custody events, employment records or commercial transactions.

Recommendation: Include an optional interoperability mechanism through which an external credential can reference:

  1. The authoritative registrar identified through GRID.
  2. The registration or credential type.
  3. The jurisdiction.
  4. The verification method.
  5. The relevant registration status.
  6. The date and time at which the verification was performed.

GRID would remain a directory of authoritative registrars, while UNTP or compatible external systems would record the transaction, asset or event.

Rationale: Identifying the competent authority is the first layer of trust. Economic use also requires a reliable connection between that authority and the specific right, asset or event being asserted.

Discussion​

[2026-09-14] John Phillips: This comment references the "About" section of the project pages - which clearly describes itself as not part of the public review. However, this section should be good enough not to create any misunderstanding and to point interested readers to more detailed deliverable content.

The observation is somewhat difficult to interpret "However, the documentation could explain more explicitly how a verified registrar identity may be referenced by external credentials representing assets, rights, invoices, custody events, employment records or commercial transactions"

This appears to imply that the reviewer is expecting that other "external" credentials might reference the verified registrar. Our logic is that the registrar recognises who it chooses to recognise and in doing so provides a trustworthy registration. Our model doesn't anticipate other parties referencing the registrar without the registrar's knowledge.

If this external unknown referencing is the intent, then the suggestion is problematic since allowing any third party to claim some relationship with the registrar that the registrar is unaware of would create a risk of fake claims.

The recommendation section is more explicit in that it elaborates the process that the reviewer has in mind. Here it seems that they are describing elements that we define in the documentation regarding the GRID registration process, the details recorded and presented by the GRID and, where applicable, the issuance of a DIA by the GRID.

My reading is that this observation and recommendation are already met by the existing documentation in the deliverable documentation set. Specifically:

  • GRID Legal Governance and Target Operating Model
  • DIA - Legal Meaning and Data Structure
  • GRID Data Schema

Issue #83: Public Review Comments - José Manuel Fernández Núñez​

Status​

State: closed | Author: John Phillips Labels: History, ODP::Stage 4 Public Review, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/83

Original Comment​

Preamble

These are review comments from José Manuel Fernández Núñez received via our groups.io email late on Sept 11 (AEST) and approved Sept 14 morning (AEST).

The email used by José was "admin@abcxsoft.com" rather than a personal email. "abcxsoft.com" is registered in the Dominican Republic. The landing page advertises a time and attendance tracking software product. This is somewhat disconcerting, however the comments appear considered and well intentioned and hence have been brought into the review process.

The comments below have been copied and pasted from the email. Some formatting has been applied to retain the visual layout of the email.

As part of processing, the comments will be split into one child task for each unique comment.


Dear UN/CEFACT GRID Project Team,

I am submitting the following comments concerning the Stage 4 Public Review of the Global Registry Information Directory.

Reviewer: José Manuel Fernández Núñez Country: Dominican Republic Professional background: Jurist, author, software developer and researcher on monetary law, verifiable economic records and the Global Truth conceptual framework.

COMMENT 1 — INTEROPERABILITY WITH TRANSACTION AND ASSET CREDENTIALS​

Page URL: https://grid.unece.org/docs/draft/About/

Issue or observation: GRID correctly limits itself to identifying authoritative registrars and does not copy the contents of national registers. However, the documentation could explain more explicitly how a verified registrar identity may be referenced by external credentials representing assets, rights, invoices, custody events, employment records or commercial transactions.

Recommendation: Include an optional interoperability mechanism through which an external credential can reference:

  1. The authoritative registrar identified through GRID.
  2. The registration or credential type.
  3. The jurisdiction.
  4. The verification method.
  5. The relevant registration status.
  6. The date and time at which the verification was performed.

GRID would remain a directory of authoritative registrars, while UNTP or compatible external systems would record the transaction, asset or event.

Rationale: Identifying the competent authority is the first layer of trust. Economic use also requires a reliable connection between that authority and the specific right, asset or event being asserted.

COMMENT 2 — STATUS, REVOCATION AND HISTORICAL EVIDENCE​

Page URL: https://grid.unece.org/docs/draft/Documents/

Issue or observation: A registrar’s authority, jurisdiction, verification service or participation status can change. A verification that is valid today may no longer be valid when a transaction is audited or disputed in the future.

Recommendation: The GRID data model and implementation guidance should clearly support:

  1. Effective dates.
  2. Expiration dates.
  3. Suspension or revocation status.
  4. Reasons for status changes, where legally publishable.
  5. Version history.
  6. A method for proving what GRID reported at a particular historical date.

Rationale: Cross-border legal certainty requires more than current-status verification. Courts, auditors, financial institutions and supply-chain participants may need reliable historical evidence showing that a registrar was recognized when a credential or transaction was issued.

COMMENT 3 — DISTINGUISHING REGISTRATION FROM ECONOMIC EVIDENCE​

Page URL: https://grid.unece.org/docs/draft/About/

Issue or observation:

Verification that a registrar is authoritative does not, by itself, establish the current existence, value, availability, ownership, absence of encumbrances or physical custody of an underlying asset.

Recommendation:

Add a clear explanatory distinction among:

  1. Verification of the registrar’s authority.
  2. Verification of an identifier or registration.
  3. Verification of legal ownership or entitlement.
  4. Verification of physical existence or custody.
  5. Verification of valuation.
  6. Verification of commercial availability or deliverability.

External credentials should identify which of these propositions are actually being certified and which remain outside their scope.

Rationale:

This distinction would reduce the risk that users treat a valid digital credential as proof of facts that the issuing registrar did not verify.

COMMENT 4 — ASSURANCE AND CERTIFIER CAPACITY​

Page URL: https://grid.unece.org/docs/draft/Documents/

Issue or observation:

Registrations and related credentials may be based on materially different assurance models, including self-declaration, administrative review, second-party assessment, independent inspection or accredited third-party certification.

Recommendation: Provide optional metadata or interoperability guidance indicating:

  1. The type of assurance.
  2. The evidence reviewed.
  3. The independence of the verifier.
  4. The verifier’s accreditation or legal authority.
  5. Whether physical inspection occurred.
  6. Whether the credential is subject to continuing monitoring.

Rationale: Making assurance levels visible would allow relying parties to distinguish an authoritative registration from an independently verified economic or physical fact.

COMMENT 5 — DISPUTES AND CONFLICTING REGISTRATIONS​

Page URL: https://grid.unece.org/docs/draft/Documents/

Issue or observation:

Cross-border transactions may encounter disputed, duplicated or conflicting registrations, particularly concerning ownership, secured interests, intellectual property, legal representation or corporate authority.

Recommendation:

The implementation guidelines should define how GRID indicates that:

  1. A registration or registrar status is disputed.
  2. Multiple authorities claim overlapping competence.
  3. An appeal or administrative review is pending.
  4. A judicial or arbitral decision has modified the legal status.
  5. Users must consult the authoritative national system for final legal effect.

Rationale: A trustworthy global directory should communicate not only positive recognition but also uncertainty, conflict and the limits of reliance.

CONCLUDING OBSERVATION​

These recommendations are compatible with GRID’s stated lightweight and decentralized approach. They do not propose turning GRID into a central registry of assets or transactions. Their purpose is to provide interoperable trust links between authoritative national registrars and external systems that certify economic facts, rights, custody, provenance, labour, contractual performance and settlement. This layered approach may help GRID become a foundational trust component for global trade, digital public infrastructure and future systems of verifiable economic records.

Respectfully, José Manuel Fernández Núñez Dominican Republic

Discussion​

[No user comments on this issue yet]

Issue #81: Consider seeding the GRID with known registers as Profile A entries​

State: opened | Author: Steve CAPELL Labels: public-review-comment::1-discussion URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/81

Original Comment​

for example from this list. grid-government-registrars.json

Discussion​

[2026-08-20] John Phillips: Do you have a specific meaning in mind for "Profile A" entries?

Assuming that it means some form of entry in the grid, then I think we can seed the grid for those that have officially confirmed their interest in the GRID (we can simply look up their official registry information and populate the GRID)

For those that we are in conversation with but have yet to confirm participation officially, then I think we need to before including them.

Does this make sense @steve.capell ?

[2026-08-20] Steve CAPELL: I do have a very specific meaning - as defined on your own website - https://grid.unece.org/docs/draft/Documents/Implementation_Guidelines?_highlight=profile#41-profile-a---grid-listing-only

It's a bit buried in the site. But maybe it's something to lift to the home-page and market more thoroughly - namely that a registrar can put an entry into the GRID with no changes - just formally advising through the member state head of delegation to UN governance that they exist.

So A -> D becomes a maturity pathway with a very low bar to start.

[2026-08-24] John Phillips: Ahh yes - hoisted by my own petard (not for the first time!)

In that sense the response is an enthusiastic 'yes' - and my clarification still holds I think - we can only include data for registrars that have agreed to participate. We can help them populate the GRID, but they need to say 'yes' for us to do so and to confirm their entry (or entries).

Same page now?

Issue #80: Alignment with W3C Recognised-Entities specification​

State: opened | Author: Steve CAPELL Labels: public-review-comment::1-discussion URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/80

Original Comment​

The W3C has produced a candidate recommendation called recognised entities - see https://www.w3.org/TR/vc-recognized-entities-1.0/. It serves the same purpose as UNTP Digital Identity Anchor (DIA) specification - see https://untp.unece.org/docs/specification/DigitalIdentityAnchor. And in the interests of hramonisation and minimising duplication for implementers, the UNTP team will update the DIA specification to align with W3C recognised entities (essentially as a conforming provide of the W3C spec).

The GRID depends on UNTP DIA so this public review comment is to ensure that the GRID team do an impact assessment of the this UNTP change (may be minimal impact on GRID because most of the change is on UNTP side).

Note that the W3C specification references GRID - https://www.w3.org/TR/vc-recognized-entities-1.0/#un-grid-cross-border-trade-example

Discussion​

[2026-08-17] John Phillips: Hi Steve,

As part of the original project brief, the published GRID documents contain a review of the UNTP DIA and suggestions on changes. Given as you say that the method/manner of harmonisation will mostly fall on the UNTP team, how do you suggest we review to a future possible but yet to be defined version?

Perhaps the simplest path is that the UNTP team take on board the existing GRID proposal for what is/was the DIA and consider them for the harmonised W3C Recognized Entities version?

Appreciate your thoughts.

[2026-08-17] Steve CAPELL: UNTP should have updated DIA to align with W3C Recognised entity spec before GRID completes public review. As you say, it's mostly UNTP impact but this ticket is really just a memory jog to make sure that, post poblic review, GRID takes a quick look at what, if anything, is needed to ensure alignment with an updated UNTP DIA.

Issue #79: Public Review - verifiable closure over material trust dependencies (Manu Fontaine)​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/79

Original Comment​

From Manu Fontaine 2026-08-16 15:29 AEST - Original received via email to the project groups.io account: https://groups.io/g/UNCEFACTGlobalTrustRegistry/topic/gtr_public_review_feedback/120772788

The email message format has been replicated below to match the original, including the highlighting of text using bold.


Dear GRID Public Review Team,

Thank you for the opportunity to comment on the GRID Public Review.

I strongly support GRID's stated objective of establishing globally interoperable digital trust infrastructure. My concern is that the proposed reliance on Verifiable Credentials as a foundational mechanism is architecturally incompatible with that objective.

I have attached a longer technical note developing the argument and recommendations in detail.

Page/Section being reviewed:​

Primary: Annex A, Technical Reference for GRID & DIA, Digital Identity Anchor, particularly "Why a credential model" and "Standards and formats," including the statement:

"The key requirement is: issuer authenticity + tamper evidence + machine checkability + revocation."

Related: Implementation Guidelines, Section 1.6, Technological Neutrality Principle; and UNTP Transparency Graphs, particularly "Verifying the Graph" and "Trust chains."

Issue:​

The stated requirement defines credential verification, not global trust.

Verifiable Credentials cryptographically establish a narrow proposition: that a particular key signed a particular assertion and that the assertion has not subsequently been altered. The material dependencies required for a relying party to determine whether that assertion should actually be trusted remain outside the VC verification boundary.

Those dependencies include, among others, the issuer's identity and authority, scope of authority, key custody and control, issuance processes, source-information provenance and completeness, software integrity, execution integrity, verification procedures, identifier resolution, status infrastructure, governance, policy, delegation, and temporal state.

Each of those dependencies may itself depend on further people, organizations, information, software, infrastructure, keys, processes, or assertions.

This creates a fundamental architectural problem.

Every additional VC can close one narrow cryptographic question while introducing another set of material trust dependencies that cannot be verified through the same mechanism. Chaining VCs therefore does not produce closure over trust. It expands a graph of cryptographically attributable assertions while simultaneously expanding the unresolved assurance graph beneath those assertions.

The more extensively VCs are composed, the more external trust dependencies can accumulate.

A graph of cryptographically valid VCs is therefore not a graph of cryptographically verified trust.

This is particularly problematic because VCs look like a solution to the trust problem. Their signatures, standardized formats, issuer identifiers, revocation mechanisms, and machine-verifiable proofs create the appearance of an end-to-end assurance architecture. In fact, the material dependencies that determine whether the assertions should be trusted remain outside that architecture.

Selecting VCs as the foundational trust mechanism therefore risks foreclosing the more fundamental architectural question that GRID should be solving.

Proposed Solution:​

UNECE should remove Verifiable Credentials as the foundational recommendation for achieving global digital trust and instead define the required assurance property first.

That property should be verifiable closure over material trust dependencies.

For any assertion used in a relying-party decision, every material dependency required to establish its provenance, authority, integrity, context, scope, temporal validity, relevant completeness, and production or verification process should either:

(a) be recursively represented and verifiable through the assurance architecture; or

(b) remain explicitly visible as an unresolved assumption or trust anchor intentionally selected by the relying party.

The architecture should preserve the ability to follow the question "why should I trust this?" recursively across the dependencies of the assertion rather than terminating that inquiry at the validity of a signed credential.

This does not require UNECE to prescribe a single national technology stack. It requires UNECE to define the functional assurance property that any implementation must be capable of satisfying if it is to participate in a genuine global trust infrastructure.

Participating jurisdictions can remain free to choose their technologies, architectures, governance mechanisms, and maturity paths. What should not vary is whether material trust dependencies are visible and verifiable enough for relying parties to make their own trust decisions.

Rationale:​

GRID's stated ambition is much larger than credential interoperability.

It seeks to support global digital trust infrastructure, cross-border verification, legal identity, transparent governance, resilient institutions, supply-chain transparency, and the objectives of the UN Global Digital Compact and 2030 Agenda.

Those objectives require the infrastructure to support an optimal trust decision by the relying party, not merely successful verification of signatures on portable assertions.

If a relying party can verify that a credential was signed by a key associated with a recognized issuer, but must still separately determine whether the issuer acted within scope, whether the source information was correct and sufficiently complete, whether the key was properly controlled, whether the software and execution environment were trustworthy, whether the verification procedure actually occurred, and whether those facts held at the relevant time, then the underlying trust problem remains unresolved.

The credential has become portable. The trust decision has not.

This also directly affects the economic case for GRID. The major economic opportunity is not simply cheaper credential checking. It is the ability to perform assurance work once and reuse it across many downstream trust decisions because the provenance and material dependencies of that assurance remain independently verifiable.

Without that property, relying parties must continue performing additional diligence outside the credential graph. The infrastructure may reduce the cost of document authentication while leaving much of the actual cost and friction of trust intact.

National sovereignty does not alter this architectural requirement.

There is an important distinction between prescribing how a country implements its domestic infrastructure and defining what properties the global infrastructure must have in order to achieve its stated purpose.

GRID's own Technological Neutrality Principle recognizes this distinction by allowing common requirements where necessary to support interoperability, integrity, authenticity, or trust.

Verifiable closure over material trust dependencies is such a requirement.

UNECE therefore does not need to mandate one implementation across participating countries. It does, however, need to decide what architecture is capable of achieving the global objective it has set.

If the recommendation is constrained to mechanisms that participating countries already use or are presently comfortable adopting, the result risks becoming a least-common-denominator interoperability framework rather than a genuine global trust infrastructure.

The decision to recommend VCs is especially consequential because it can make the architecture appear further along than it is. VCs provide visible cryptographic proofs, interoperable formats, and readily demonstrable verification flows. But the unresolved trust dependencies remain outside those proofs.

The result is a system that can become increasingly effective at proving who signed assertions while never acquiring the architectural property required to determine why those assertions should be trusted.

For GRID's stated objectives, that distinction is foundational.

I therefore recommend that UNECE reconsider the VC-centric architectural starting point and first establish verifiable closure over material trust dependencies as the governing requirement against which candidate technologies and national implementations are evaluated.

The attached technical comment develops this argument in detail.GRID_Public_Review_Verifiable_Closure.pdf

Discussion​

[2026-08-17] Sankarshan Mukhopadhyay: [updated the original comment with more context/color]

Initial observation: I think this feedback usefully highlights the distinction between verifying a credential and deciding whether to trust the claim it carries.

Where I differ is with the conclusion that this makes Verifiable Credentials inherently incompatible with GRID. I do not read GRID as claiming that successful verification of a VC is, by itself, sufficient assurance. A verifier may still need to understand who made the claim, their authority and scope, the provenance of the underlying information, its lifecycle and status, and the governance under which it was issued.

The architectural requirement, therefore, should be that these material sources of assurance and remaining trust assumptions are explicit and can be examined where necessary. The credential format is only one part of that architecture.

In that sense, I find “verifiable closure” useful as a test: not as an argument against VCs, but as a way of asking whether GRID exposes enough of the assurance basis for a relying party to make its own trust decision.

[2026-08-17] John Phillips: Manu's comments deserve a considered response - a lot of effort and thinking went into their creation. These then are some initial comments from me and are not final in any sense.

I agree with @sankarshanmukhopadhyay's comments above. I also think that the comments evidence an interesting blurring of the scope of the GRID project and proposed entity with the other UN/CEFACT projects (UNTP, UNVTD etc.). The scope of the GRID should be clear - we should check that this is the case.

As an example of scope - not all registries listed in the GRID will be digitally enabled. We do not (cannot) expect all Registrars in all UN Member States to have the same level of digital maturity. This means that not all listed Registrars in the GRID will be digitally enabled. While we are duty bound to encourage a standardized approach (which in the current documentation means the use of a VC, DID, and DIA), some Registry's already have their own approach, and some may not be capable of a self-sovereign digital capability. The GRID must be able to list those registrars who use different methods as well as those who don't (yet) publish digital artefacts.

This means that not every participating member in GRID needs to be digitally enabled. They won't all issue DIAs, nor can GRID issue a DIA recognising them if they cannot create and control a DID.

GRID needs to be useful in a world where not every interaction is digital.

The decision to "trust" is a context based decision individually made by a party. UNTP and GRID make no claims to understand all possible contexts or uses of supply chain information. Trust implies sufficient confidence to take action in the presence of uncertainty. We don't need "trust" if we have perfect information and absolute certainty about the effect of our action. However, GRID and the other UNTP related projects make no claims to provide perfect information or to "solve" for trust. Our aim is to provide a more transparent, verifiable basis on which a decision can be made. At a simple level ,we want to improve the existing trade document exchange by making these documents "verifiable" in that the receiving party can check who issued them and that they haven't been tampered with. That alone is a significant improvement on existing paper based trade document exchange.

Put simply: GRID does not seek to tell any party to trust something or not. UN Member States may trust other UN Member States, or not. Some UN Member States may adopt equivalent laws or have mutual recognition, others may not. GRID imposes no requirement on trust, it seeks to provide trustworthy information about which Registrars manage which Registers in which UN Member States.

This goes to a broader observation: in a real-world human geo-political system, technology alone can never determine whether something should be trusted. There are many (many) other factors that determine whether something might be trustworthy. Trust over IP summarises this in two stacks: technology and governance. Others might want to cite the legal framework as well as governance framework, as the underlying incentive model.

This helps to explain the specific and deliberately limited context in which the GRID makes use of the W3C VC data model and DID's. In the context of the GRID's use, they provide a foundation that is used by issuing parties to make cryptographically verifiable claims. The "verifiable" quality is merely that a specific issuer signed the data and that it hasn't been tampered with since it was signed. That is all we expect of them.

[2026-08-25] John Phillips: Summary of my thinking (see above comment for earlier musing)

  1. Although alluding to wider text references, the reviewer specifically refers to a section of text within Annex A and states that "The stated requirement defines credential verification, not global trust." - this is correct.
  2. The reviewer goes on to assert that, in their opinion, we should be defining a system for global trust. We make no such claim for GRID.
  3. The rest of their observations and proposed resolution hinges on their own assertion - that we should be defining a system for global trust.
  4. The scope and goal of the GRID is not to "define system for global trust", certainly not in the way that the reviewer presents it. The GRID and the family of projects started by the UNTP Specification aim to make trade documentation, and the claims within them discoverable, resolvable and verifiable. This enables the "transparency graph".
  5. GRID supports that aim by providing a trustworthy source of information about the authoritative registrars that issue registrations. It further supports that aim by issuing and encouraging GRID participants to issue UNTP compliant verifiable credentials for their registered entities.
  6. It is important that "verifiable" is understood to be mean the standard definition for its use by GRID, W3C, UNTP etc. This should be clear in our Glossary.

Proposed resolution 7. Revisit the Glossary definition to ensure it accurately reflects the meaning of "verifiable" that we intend 8. Check the use(s) of the term "verifiable" in other text to ensure that it aligns with the Glossary definition. 9. Consider if and how we might prevent others from arriving at the same mis-understanding about the scope and goals of GRID.

I understand that this is a very limited change compared to the breadth and depth of the review comment. I respect the views and ambition of the reviewer and the work they have done to build a solution (see hushmesh.com), but they have taken their mis-representation of the goal of the GRID and used that mis-representation to argue for changes that aren't warranted.

[2026-08-25] John Phillips: These changes are being put into this branch: https://opensource.unicc.org/un/unece/uncefact/gtr/-/tree/PublicReviewCommentsUpdate02

Issue #78: Align lifecycle, status and revocation across both credential layers​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/78

Original Comment​

URL page to which comment relates​

DIA Legal Meaning and Data StructureAnnex - lifecycle, key management and verification flow Pilot Design Architecture Sections 2.1, 2.4, 3.2, 3.3 and 5.1 Pilot Participant Guide Pilot Limitations - Revocation Infrastructure

Comments​

Align lifecycle, status and revocation across both credential layers The DIA Annex defines entity-level lifecycle management by the Registrar and Registrar-level key events in GRID. The pilot separately creates a GRID-issued credential for the Registrar, while the Participant Guide states that pilot credentials contain no credentialStatus block. The two credential layers therefore require separate, explicit status authorities and test coverage.

Proposed changes​

Proposed change Include a credentialStatus block for the GRID-issued Registrar credential. If this remains deferred during the pilot, document the reason for the deferral, the interim relying-party guidance, how suspension or withdrawal is communicated, and the limitations this places on conclusions drawn from pilot verification results. Add a pilot lifecycle and exception test matrix covering Registrar suspension and withdrawal; Registrar key rotation and compromise; entity dissolution or insolvency; subject-key replacement; credential supersession; status-endpoint failure; and historical verification after change. For each credential layer, define the issuer, status authority, credentialStatus method, revocation or suspension event, recovery procedure and relying-party result.

Discussion​

[2026-07-26] John Phillips: Initial comments

The Pilot design was not initially intended to be the operational design, although as the design matures it becomes more operationally capable. The initial decision not to use credentialStatus was a simplification for the Pilot participants and the GRID Pilot implementation itself.

Work has since been done to consider how best to manage lifecycle status over the years that we would expect the recognition to be important. While credentialStatus is important in this regard, it isn't sufficient on its own to manage the historical status of recognition at a past point in time.

The decision on this issue will be related to #75

Issue #77: Organisational identity does not establish authority to act​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/77

Original Comment​

URL page to which comment relates​

DIA Legal Meaning and Data Structure Section 2.1; Article I; Article III

Comments​

Organisational identity does not establish authority to act A DIA binds an authoritative organisational registration to a subject-controlled DID. It does not by itself establish the identity or authority of an individual acting for the organisation, a regulated licence, or permission to perform a specific transaction. In customs, logistics, banking and electronic transferable record workflows, relying parties could otherwise conflate organisational identity with authority to submit, approve, sign, transfer control or bind the organisation.

Proposed changes​

Proposed clause under Article I or Article III 'A DIA attests to registered organisational identity and to the Registrar's verification of control of the associated identifier at the time the linkage was checked for issuance. It does not, by itself, attest to the identity, role, employment status, professional or regulatory qualification, legal authority, delegated authority, or transaction-specific permission of any individual acting for the organisation. Those matters require separate authoritative credentials, mandates or verification procedures.' Add an implementation note showing how DIA verification should be combined with representative identity, role or authority, and transaction-permission evidence.

Discussion​

[2026-07-26] John Phillips:

Consideration

Section 2.1 introduces the idea that data that is issued by and controlled by a Registrar has a different quality to data that the registrar receives and checks, and from data that the Registrar receives but does not (or cannot) check.

This idea is continued in section 2.2. and 2.3 and in section 2.3 a proposed approach is recommended where the Registrar can indicate the type of data in the DIA against three criteria:

  • Authoritative (they issue and control this data)
  • Verified (they received and verified this data, the subject's DID is an important example of this)
  • Asserted (the subject provides/asserts this data but it is not checked by the Registar)

There are historical references to "GTR" in these sections that should be replaced with "GRID"

The central idea in the comment is not accepted. The GRID recognises that individual UN Member States are sovereign. They have their own legal framework and to be recognised as an organisation in one jurisdiction may demand obligations and duties and provide rights that are not the same in another jurisdiction.

Further, the GRID is focused on DIA issuance by Authoritative Registrars. These DIAs have a significant role to play in trustworthy trade, but it is recognised that there are many other trade documents required to support any trade transaction. The broader consideration of roles, authority and transaction-permission is provided by the full suite of the UNTP specification.

Proposed changes:

No changes to Sections 2.1, 2.2, 2.3 or Articles I and III EXCEPT for the replacement of "GTR" references with "GRID".

NOTE : Commit https://opensource.unicc.org/un/unece/uncefact/gtr/-/commit/a6068516a6487cdaf43955afc98311fbc451d3c1 changed GTR references to GRID.

[2026-07-30] John Phillips: These changes have already been made (see commit link above) and the changes visible in the main branch - this comment is proposed to be ready-for-review.

[2026-08-24] John Phillips: Closed after merge 43

https://opensource.unicc.org/un/unece/uncefact/gtr/-/merge_requests/43

Issue #76: Use of “not verifiable through GRID” rather than a universal “Unverified” designation​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/76

Original Comment​

URL page to which comment relates​

DIA Legal Meaning and Data Structure Section 1.1; Article II.2; Article IV

Comments​

Use 'not verifiable through GRID' rather than a universal 'Unverified' label Section 1.1 states that DIA issuance is not limited to GRID participants and Article IV makes participation voluntary. Article II.2 nevertheless deems a DIA from an entity not listed in GRID 'Unverified' for international trade facilitation. That term may be read as a general judgement on the credential rather than a statement about GRID's own verification scope, and may unintentionally imply that GRID is the exclusive trust framework.

Proposed changes​

Proposed replacement for Article II.2 'The GRID serves as a trusted discovery and verification directory for participating Authoritative Registrars. A DIA issued by an entity not listed in the GRID is not verifiable through GRID. Its validity, reliability or recognition may be established under another applicable legal, governmental, standards-based or recognised trust framework. Relying parties remain responsible for applying their applicable trust policy.'

Discussion​

[2026-07-26] John Phillips: Initial thoughts.

Agree with the observation here that the wording of Article II is in conflict with the idea that any organisation can choose to issue a UNTP DIA (which is stated in Section 1.1). Article IV is unrelated to comment and it isn't clear why it is included in the observation.

Section 1.1 is correct in that the GRID cannot and shouldn't not prevent the adoption and use of UNTP constructs including the DIA. Indeed, we should encourage the wider use of the UNTP specification.

The MEANING (legal and otherwise) of any issued credential is inherited from the Issuer and the legal framework in which they are recognised and the authority which they grant the DIA. As stated elsewhere and in alignment with MLETR, the requirement is that the electronic version of a record is legally equivalent to the physical version.

The primary suggestion of the comment is reasonable.

First reading is the Article II will need rewording.

@ANICA ?

[2026-07-30] John Phillips: Adopted with minor changes:

The GRID serves as a trusted discovery and verification directory for participating Authoritative Registrars. A DIA issued by an entity not listed in the GRID is not verifiable through the GRID. Establishing the validity of such a non-GRID verifiable DIA is the responsibility of the relying party.

Updates proposed in branch PublicReviewComments01

[2026-08-24] John Phillips: Closed after merge !43

Issue #75: Historical and as-of verification after withdrawal​

Status​

State: closed | Author: John Phillips Labels: History, public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/75

Original Comment​

URL page to which comment relates​

GRID Legal Governance and Target Operating Model Appendix A.4 - Termination and Withdrawal

Comments​

Historical and as-of verification after withdrawal Appendix A.4.2 states that, after withdrawal, the Registrar's metadata and public keys are removed from active discovery and previously issued DIAs may no longer be discoverable or verified through GRID. Trade records may require later verification for audit, financing, insurance, regulatory review or dispute resolution. Removing the active entry without preserving time-specific evidence may prevent a relying party from proving that the Registrar and signing key were valid when a transaction occurred.

Proposed Changes​

Proposed change Maintain a tamper-evident, versioned historical verification archive separate from active discovery. Preserve prior Registrar metadata, public keys, key status, effective dates, suspension and withdrawal events, and relevant status-list versions. Provide an 'as-of' verification method that can confirm the Registrar and key status at a specified transaction time. Clearly label withdrawn or superseded entries as historical and prevent their use for current trust decisions.

Discussion​

[2026-07-26] John Phillips: Initial thoughts (not yet conclusive)

The intent of this section is to ensure that the GRID has no overarching authority over participants. If a participant chooses to withdraw, then the GRID must accept that decision.

The observation raises an interesting point about whether historical participation in the GRID should be a matter of record for the GRID to maintain.

One of the key principles proposed for the GRID is that it make no change to the existing legal and commercial approach adopted by its members. Further, the design was originally considered to provide the current state, rather than past state of participation.

There is a subsequent impact on this consideration: if the GRID has issued DIAs for the withdrawing participant, should it delete them? The answer to this question feels more likely to be "no", and this would tend to the proposed approach.

There is work being developed to consider how lifecycle events (history of status changes) should be managed DIA issuance.

Work still to be done on considering how best to address these comments.

[2026-07-27] Sankarshan Mukhopadhyay: While considering this issue, the project will need to define the following:

  1. Retention: How long GRID historical records and retired verification material are retained, and whether trade, audit, limitation, dispute, or regulatory periods require differentiated retention rules.
  2. Withdrawal versus erasure: Which information constitutes a record of GRID’s own past actions and assertions and may therefore remain retainable, and which participant-supplied information must be removed, minimised, or access-restricted.
  3. Key compromise: Whether a compromise affects only uses after the compromise became effective, or whether a competent authority may declare an earlier effective time. Any retroactive effect must be explicit, attributable to an authorised decision-maker, and supported by evidence. It must not be inferred merely from the current status of the key.
  4. Evidence integrity: Whether historical records require append-only versioning, signed snapshots, transparency-log-style commitments, or another mechanism capable of detecting backdating, selective rewriting, or operator equivocation.
  5. Availability: Whether historical verification is publicly available, access-controlled, or supplied only for defined audit, regulatory, dispute-resolution, or legal purposes.
  6. (Optional) Protocol alignment: Whether GRID exposes this capability directly, through a GRID profile of TRQP, or through another interface. TRQP alignment would be useful, but the underlying governance obligation should not depend on the completion or adoption of the current RFE.

My sense is that this issue should be addressed through coordinated changes to:

  • Appendix A.4, so that the legal effect of withdrawal distinguishes active delisting from historical retention;
  • the DIA operational guidance, so that verification explicitly selects the key and Registrar state applicable at issuance or transaction time rather than relying only on current discovery;
  • the GRID technical reference and data model, so that verification material, participation status, and relevant lifecycle events can be versioned and associated with effective periods;
  • operating procedures covering retention, access, key compromise, exceptional retroactive decisions, and evidentiary use; and
  • pilot test cases covering withdrawal, key rotation, suspension, compromise, cancellation, and historical DIA verification.

This would preserve the core principle that GRID cannot prevent a participant from leaving, while avoiding a design in which withdrawal unintentionally destroys the evidence required to verify legitimate historical trade activity.

Suggested amendment to Appendix A.4.2​

Effect of Withdrawal: Upon the effective date of withdrawal, the Registrar’s metadata, service endpoints, and verification material shall be removed from, or marked inactive within, the active GRID discovery layer. They shall not be returned or interpreted as supporting a current trust decision.

GRID shall retain, in a logically separate historical verification record, the minimum information necessary to establish the Registrar’s participation status and applicable verification material for periods before withdrawal. Historical records shall be clearly identified as inactive or withdrawn, shall include their applicable effective periods and relevant lifecycle events, and shall not be presented as evidence of current participation, recognition, or authority.

GRID shall define an “as-of” verification capability through which an authorised relying party can determine the Registrar and key status applicable at a specified time and identify the record, version, or evidence basis used to produce that determination.

The applicable retention period, access conditions, evidentiary limitations, privacy and minimisation controls, and treatment of exceptional events such as key compromise, cancellation, or retroactive status changes shall be defined through GRID governance and operating procedures.

Minimum historical record​

The historical record retained by GRID should include, at minimum:

  • the stable Registrar or GRID record identifier;
  • the participation state and its effective period;
  • the verification material or key identifier applicable during that period;
  • key activation, rotation, suspension, compromise, retirement, and withdrawal timestamps, where applicable;
  • the source governance event and reason code, preserving distinctions such as voluntary withdrawal, suspension, authority cancellation, and key compromise;
  • the identifier of each authoritative record or version;
  • predecessor or successor relationships where lifecycle changes are represented through immutable successor records;
  • the resolution rule or lifecycle profile applicable to the record; and
  • a reference to supporting evidence or provenance sufficient to make an as-of determination auditable.

For each as-of response, GRID should additionally return:

  • the requested time;
  • the evaluation time;
  • the record or version selected;
  • the resolution method applied; and
  • an evidence or provenance reference.

The historical archive should retain only the minimum metadata necessary to support GRID verification and accountability. It should not become a copy of national registry content, participant business records, or the underlying entity register.

As-of response semantics​

For an as-of query, GRID should return the state applicable at the requested time. Any events learned or occurring after that time should be represented separately from the historical snapshot.

For example:

{
"registrarId": "grid:registrar:example",
"timeRequested": "2025-11-15T00:00:00Z",
"timeEvaluated": "2026-07-27T01:30:00Z",
"statusAtRequestedTime": "active",
"verificationMaterialAtRequestedTime": {
"keyId": "grid:registrar:example#key-2025",
"validFrom": "2025-01-01T00:00:00Z",
"validUntil": "2026-03-01T14:30:00Z"
},
"selectedRecordId": "grid:history:example:2025-01",
"currentStatus": "withdrawn",
"currentStatusAsOf": "2026-07-01T00:00:00Z",
"sourceStatus": "voluntary-withdrawal",
"resolutionProfile": "https://example.org/grid/profiles/historical-resolution/v1",
"evidenceReference": "https://example.org/grid/history/records/..."
}

A historical response must not be usable, without explicit temporal interpretation, as evidence that the Registrar is currently recognised or authorised.

[2026-08-04] Jo Spencer: Initial thoughts... What constitutes a valid presentation of VCs from authoritative sources is not for the GRID to define. The operational governance of trust ecosystems sets these rules by defining

  • which registry issues the DIA for the credential issuer
  • which types, versions and timeframes for credentials are requred
  • how the claim content of these credentials are assessed and combined The content that has been identified above is required to do this in the GRID context. If a "current" DIA (or a DIA of specific versions) is required, re-accreditation will be required and these DIA versions will only be accepted. The GRID cannot be a dependency on the use of these DIAs, but reaccreditation can be demanded by the involved Registry/Registrar and they can demand revocation of GRID DIAs as a result.

[2026-08-11] Sankarshan Mukhopadhyay: Thinking over our discussion of this topic at the last meeting I'd like to include one additional distinction that may be useful when resolving this issue.

GRID's responsibility should be to make historical participation and verification evidence deterministically resolvable, rather than to determine whether a historical DIA or credential is acceptable for a particular transaction or trust ecosystem. The latter remains a matter for the applicable operational governance framework, which may require current accreditation, particular DIA versions, particular credential versions, or other conditions.

In practical terms, an as-of capability should therefore establish something closer to:

"According to the authoritative GRID history and the applicable resolution rule, this was the Registrar/key/status applicable at time T."

It should not imply:

"This historical evidence is acceptable for the relying party's present trust decision."

It would also be useful to keep three things machine-readably distinct:

  • the state applicable at the requested historical time;
  • the current state at the time the query is evaluated; and
  • later-known events that affect interpretation of the historical result.

This avoids conflating historical verification with current recognition and reinforces the separation @JoatUNICC Jo identifies above: GRID supplies authoritative, temporally qualified evidence; the consuming governance framework determines the admissibility and trust consequence of that evidence.

[2026-08-11] Jo Spencer: I agree with your suggestion @sankarshanmukhopadhyay. I hope you don't mind my re-interpretation, below.

The GRID (or potentially a GRID instance) has to be able to stand behind its issued DIAs. The verification process must be able to show that (a) the verification status of the presented DIA at the specified time (b) the specific status of the DIA at the time stipulated.

Whilst it would be good to retain an audit of "later-known event that affect interpretation of this historical result" this may be difficult to define and design for, let alone provide an interpretable basis for automated assessment of the current status of historical DIAs. This may be an option, but not mandated.

[2026-08-25] John Phillips: I think it is worth recognising that there are three sets DIA issuers that we are interested in:

  1. the DIA's issued and managed by the GRID in recognition of participating members entries, and
  2. the DIA's issued by the GRID Members (the Participating UN Member State Registrars and their Registries)
  3. the DIA's that may be issued by any other parties following the UNTP Specification.

For 1. we need to define what the GRID will do. For 2. we can recommend what GRID members do, but we have no authority to force compliance. For 3. we could offer general guidance, but similarly to 2., we have no authority over how other parties may choose to use the DIA.

To response to the original review comment, our primary focus should be on 1. - what the GRID will do. This should serve as guidance for others, so we give guidance to 2 and 3 by focusing on 1.

For the sake of clarity, "the GRID" refers to the UN/CEFACT operated GRID as defined in our documentation, not any other "GRID pattern following instance".

While we are of course interested in the general trade documentation scenario, this is not in scope for this project. We are only focused on the GRID operation and its use of the DIA. Our considerations on the wider trade implications will hopefully be useful for others who do have this scope, but, strictly speaking, they are not our problem to solve.

[2026-10-15] John Phillips: Agreed as important and requiring changes.

Addressed by new content addressing verifiable history. See Merge !68

Issue #74: Credential-layer and issuer ambiguity​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::2-ready-for-review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/74

Original Comment​

URL page to which comment relates​

DIA Legal Meaning and Data Structure Section 1.1; Article II.1 GRID Legal Governance and Target Operating Model Sections 1, 3.2 and 7.2; Appendix A.1 Pilot Design Architecture Overview; Sections 2.1, 2.4, 3.2, 3.3 and 5.1

Comments​

Credential-layer and issuer ambiguity The draft presents three inconsistent issuance models. Section 1.1 states that organisations other than GRID-participating Authoritative Registrars may choose to issue DIAs. Article II.1 provides that a DIA may only be issued by an Authoritative Registrar with a statutory mandate to maintain a register. The Pilot Design has the GRID Hub issue a UNTP DIA to a Registrar using the GRID signing key. The Governance Framework describes GRID as a discovery and reference service rather than an Authoritative Registrar. Unless these are intended as distinct credential classes, the scope statement, normative issuance rule and pilot architecture are inconsistent. This also creates ambiguity concerning legal meaning, liability, status authority and revocation.

Proposed change​

Reconcile Section 1.1 and Article II.1 by defining who may issue a DIA under the UNTP specification and under what conditions. Reserve 'Digital Identity Anchor' for the Registrar-to-registered-entity credential. Define a separate credential, such as 'GRID Registrar Trust Credential' or 'GRID Participation Credential', for the GRID-to-Registrar relationship. Add a normative trust-chain diagram or table identifying, for each credential, the issuer, subject, claims, legal meaning, signing key, status authority, revocation mechanism and relying-party verification steps. Update the pilot terminology, schema and examples accordingly.

Discussion​

[2026-07-25] John Phillips:

Review determination​

Text should be updated, but the proposed change is rejected as it reflects a misunderstanding of the approach defined in the documents.

Reasoning​

####### "DIA - Legal Meaning and Data Structure" document.

Section 1.1 states that organisations other than GRID-participating Authoritative Registrars may choose to issue DIAs.

This is true, and correct. The full sentence paragraph is:

The issuance and use of a UNTP Specification DIA is not limited to Authoritative Registrars participating in the GRID. Other organisations may choose to issue DIAs to other organisations, and the ability to issue a DIA is not a precondition or post-expectation of participating in the GRID.

The opening sentence of this document is:

This document explores the legal nature and technical architecture of the UNTP Digital Identity Anchor (DIA) when issued by Authoritative Registrars who participate in the Global Registry Information Directory (GRID).

These sentences establish that the DIA is a UNTP specification credential, not a GRID specified credential. Therefore its use is not limited to the GRID or GRID participants. Importantly, this opening section concludes with:

This document establishes the "Functional Equivalence" of a DIA as issued by Authoritative Registrars to traditional registry extracts while providing the cryptographic verifiability required for supply chain transparency.

(emphasis added to this comment)

Thus the scope that the document defines is solely for DIA's issued by Registrars choosing to operate under the GRID scheme.

This then gives the context for the subsequent content in the document.

Proposed change is to modify the last sentence in section "1.1 Purpose and Scope" to state:

This document establishes the legal meaning of a DIA as issued by Authoritative Registrars participating in the GRID in comparison to traditional registry extracts, while providing the cryptographic verifiability required for supply chain transparency.

The key change is highlighted.

####### GRID Legal Governance and Target Operating Model

  • Section 1.: No change proposed
  • Section 3.2: No change proposed (text makes no reference to DIA issuance?)
  • Section 7.2: No change proposed (text makes no reference to DIA issuance?)

No change required to document

####### Pilot Design Architecture

Review comment raised:

The Pilot Design has the GRID Hub issue a UNTP DIA to a Registrar using the GRID signing key. The Governance Framework describes GRID as a discovery and reference service rather than an Authoritative Registrar. Unless these are intended as distinct credential classes, the scope statement, normative issuance rule and pilot architecture are inconsistent. This also creates ambiguity concerning legal meaning, liability, status authority and revocation.

Response: Disagree. There is no inconsistency with the GRID issuing a DIA for participating Registrars who have the capacity to manage a DID. The service enabled by UNTP overall (including but not limited to the use of DIAs) is Discover - Resolve - Verify (see https://untp.unece.org/docs/specification/#for-identifier-registers)

Overall conclusion​

The review comments anticipate the need to generate specific types of credentials for (what the reviewer sees as) specific uses. This is rejected.

[2026-07-30] John Phillips: Changes made to this branch https://opensource.unicc.org/un/unece/uncefact/gtr/-/tree/PublicReviewCommentsUpdate01?ref_type=heads

[2026-08-24] John Phillips: After three weeks with no comment or dissent - I will merge this change.

[2026-08-24] John Phillips: Closed on merge !43

Issue #68: Chris Papp - Set of Public Review Comments​

Status​

State: closed | Author: John Phillips Labels: public-review-comment::1-discussion URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/68

Original Comment​

Email from reviewer:​

Dear John and Alina,

Please find attached my completed comment log for the public review of the Global Registry Information Directory, GRID, Project P1138.

I am submitting these comments in my capacity as a digital trade practitioner, informed by my work developing the TransPacific Trade Nexus, TPTN, a Canada-first digital trade infrastructure initiative.

The comments address five implementation and governance matters:

  1. Credential-layer and issuer ambiguity involving GRID-issued Registrar credentials and Registrar-issued Digital Identity Anchors
  2. Historical and time-specific verification following Registrar withdrawal
  3. Use of “not verifiable through GRID” rather than a universal “Unverified” designation
  4. The distinction between organisational identity and authority to act
  5. Lifecycle, credential-status and revocation responsibilities across both credential layers

The comments are intended to support clarity in the trust model, relying-party responsibilities, interoperability and evidentiary continuity.

They do not imply GRID or UN/CEFACT endorsement of TPTN, nor any claim of GRID conformance.

Thank you for the opportunity to contribute to the review.

Kind regards,

Chris Papp Founder & CEO | TransPacific Trade Nexus (TPTN)


Word Document Review Document rendered as Markdown​

Public Review: Global Registry Information Directory (GRID) – P1138
Template for comments and observations

Please return completed template to Public Review Editors - John Phillips and Alina Nica Gales: UNCEFACTGlobalTrustRegistry@groups.io
Public review page: Public Review: Global Registry Information Directory (GRID)

Date submission:22 July 2026
Submitted byName:Chris Papp
Organization:TransPacific Trade Nexus (TPTN)
Country:Canada

Please propose suggested changes in order to make the Draft align with your comments (only those with proposed changes can be fully considered).

Ref. (leave blank)URL page to which comment relatesCommentsProposed changesWorking Group Observations (leave blank)
DIA Legal Meaning and Data Structure Section 1.1; Article II.1 GRID Legal Governance and Target Operating Model Sections 1, 3.2 and 7.2; Appendix A.1 Pilot Design Architecture Overview; Sections 2.1, 2.4, 3.2, 3.3 and 5.1Credential-layer and issuer ambiguity The draft presents three inconsistent issuance models. Section 1.1 states that organisations other than GRID-participating Authoritative Registrars may choose to issue DIAs. Article II.1 provides that a DIA may only be issued by an Authoritative Registrar with a statutory mandate to maintain a register. The Pilot Design has the GRID Hub issue a UNTP DIA to a Registrar using the GRID signing key. The Governance Framework describes GRID as a discovery and reference service rather than an Authoritative Registrar. Unless these are intended as distinct credential classes, the scope statement, normative issuance rule and pilot architecture are inconsistent. This also creates ambiguity concerning legal meaning, liability, status authority and revocation.Proposed change Reconcile Section 1.1 and Article II.1 by defining who may issue a DIA under the UNTP specification and under what conditions. Reserve 'Digital Identity Anchor' for the Registrar-to-registered-entity credential. Define a separate credential, such as 'GRID Registrar Trust Credential' or 'GRID Participation Credential', for the GRID-to-Registrar relationship. Add a normative trust-chain diagram or table identifying, for each credential, the issuer, subject, claims, legal meaning, signing key, status authority, revocation mechanism and relying-party verification steps. Update the pilot terminology, schema and examples accordingly.
GRID Legal Governance and Target Operating Model Appendix A.4 - Termination and WithdrawalHistorical and as-of verification after withdrawal Appendix A.4.2 states that, after withdrawal, the Registrar's metadata and public keys are removed from active discovery and previously issued DIAs may no longer be discoverable or verified through GRID. Trade records may require later verification for audit, financing, insurance, regulatory review or dispute resolution. Removing the active entry without preserving time-specific evidence may prevent a relying party from proving that the Registrar and signing key were valid when a transaction occurred.Proposed change Maintain a tamper-evident, versioned historical verification archive separate from active discovery. Preserve prior Registrar metadata, public keys, key status, effective dates, suspension and withdrawal events, and relevant status-list versions. Provide an 'as-of' verification method that can confirm the Registrar and key status at a specified transaction time. Clearly label withdrawn or superseded entries as historical and prevent their use for current trust decisions.
DIA Legal Meaning and Data Structure Section 1.1; Article II.2; Article IVUse 'not verifiable through GRID' rather than a universal 'Unverified' label Section 1.1 states that DIA issuance is not limited to GRID participants and Article IV makes participation voluntary. Article II.2 nevertheless deems a DIA from an entity not listed in GRID 'Unverified' for international trade facilitation. That term may be read as a general judgement on the credential rather than a statement about GRID's own verification scope, and may unintentionally imply that GRID is the exclusive trust framework.Proposed replacement for Article II.2 'The GRID serves as a trusted discovery and verification directory for participating Authoritative Registrars. A DIA issued by an entity not listed in the GRID is not verifiable through GRID. Its validity, reliability or recognition may be established under another applicable legal, governmental, standards-based or recognised trust framework. Relying parties remain responsible for applying their applicable trust policy.'
DIA Legal Meaning and Data Structure Section 2.1; Article I; Article IIIOrganisational identity does not establish authority to act A DIA binds an authoritative organisational registration to a subject-controlled DID. It does not by itself establish the identity or authority of an individual acting for the organisation, a regulated licence, or permission to perform a specific transaction. In customs, logistics, banking and electronic transferable record workflows, relying parties could otherwise conflate organisational identity with authority to submit, approve, sign, transfer control or bind the organisation.Proposed clause under Article I or Article III 'A DIA attests to registered organisational identity and to the Registrar's verification of control of the associated identifier at the time the linkage was checked for issuance. It does not, by itself, attest to the identity, role, employment status, professional or regulatory qualification, legal authority, delegated authority, or transaction-specific permission of any individual acting for the organisation. Those matters require separate authoritative credentials, mandates or verification procedures.' Add an implementation note showing how DIA verification should be combined with representative identity, role or authority, and transaction-permission evidence.
DIA Legal Meaning and Data Structure Annex - lifecycle, key management and verification flow Pilot Design Architecture Sections 2.1, 2.4, 3.2, 3.3 and 5.1 Pilot Participant Guide Pilot Limitations - Revocation InfrastructureAlign lifecycle, status and revocation across both credential layers The DIA Annex defines entity-level lifecycle management by the Registrar and Registrar-level key events in GRID. The pilot separately creates a GRID-issued credential for the Registrar, while the Participant Guide states that pilot credentials contain no credentialStatus block. The two credential layers therefore require separate, explicit status authorities and test coverage.Proposed change Include a credentialStatus block for the GRID-issued Registrar credential. If this remains deferred during the pilot, document the reason for the deferral, the interim relying-party guidance, how suspension or withdrawal is communicated, and the limitations this places on conclusions drawn from pilot verification results. Add a pilot lifecycle and exception test matrix covering Registrar suspension and withdrawal; Registrar key rotation and compromise; entity dissolution or insolvency; subject-key replacement; credential supersession; status-endpoint failure; and historical verification after change. For each credential layer, define the issuer, status authority, credentialStatus method, revocation or suspension event, recovery procedure and relying-party result.

Discussion​

[No user comments on this issue yet]

Issue #63: Prepare Site for Review​

Status​

State: closed | Author: John Phillips Labels: ODP::Stage 4 Public Review URL: https://opensource.unicc.org/un/unece/uncefact/gtr/-/issues/63

Original Comment​

We need to create a simple way for review comments to be generated and processed that is as simple as possible but also encourages structured review comments that can be shared and processed accordingly.

This means some a new page describing the review process, as well as some banner changes to alert viewers that there is a review.

Discussion​

[No user comments on this issue yet]