Session Date/Time: 29 Jan 2025 15:00

CORE

Summary

This interim meeting of the CORE Working Group focused on two main documents: "Constrained RESTful Environments (CoRE) Content-Format-based URI scheme for CoAP (CoRE-CFF-Update)" and "Concise Resource Identifiers (CoRE-HREF)". The discussion for core-cff-update revolved around addressing open GitHub issues from its Working Group Last Call (WGLC), including details on media type parameter registration, registry updates, and review processes. For core-href, the group reviewed recent changes to address long-standing issues, particularly concerning composability, integration with ACE, IDNA, and Zone Identifiers, leading towards a second WGLC. An AOB item briefly touched on the scope of DNS Update in DNS-over-CoAP (DCoAP).

Key Discussion Points

CoRE Content-Format-based URI Scheme for CoAP (CoRE-CFF-Update)

The discussion focused on several GitHub issues following the document's Working Group Last Call:

Issue 4 (IAB BIS WG Notification): Confirmed that the IAB BIS WG mailing list was included in the WGLC notification. No responses were received from that group.
Issue 22 (Media Type Parameter Quoting): Discussed the semantic equivalence of quoted/unquoted media type parameters and the need for a canonical/preferred form in registrations. The preference is for unquoted forms where allowed by encoding rules, followed by a semicolon and exactly one space. Existing entries would not be modified.
Issue 23 (Equivalent Media Type Strings): Highlighted examples of functionally equivalent media type strings due to variations like case sensitivity (e.g., in urn namespace identifiers). Suggested adding an example to the document to advise designated experts to consult parameter definitions for case sensitivity.
Issue 24 (New Registry Column and Provisional Types): Proposed adding a new column to the registry for the associated base media type, with a clickable link to its IANA assignment page. This would act as a check for existence. The concept of "provisional content formats" was discussed, noting its existing unofficial use (e.g., for application/voucher-plus-cbor). The need to formalize temporary/provisional assignments across different registry ranges (0-255, FCFS, ITF review) was raised.
Issue 26 (Lightweight Checks and Expert Review): The term "lightweight checks" was deemed problematic and should be replaced. The discussion clarified that expert review should be explicit for all registry ranges, including the ITF Review range and ISG approval, to ensure thorough review.
Issue 28 (Formalizing Temporary Note): This issue was considered covered by the discussion on Issue 24, focusing on formalizing the concept of temporary entries in the registry for both expert-reviewed ranges (0-255 and ITF Review).
Other Potential Issues: A hypothetical scenario was mentioned where an existing media type with an assigned number in the 0-255 range could also be registered by an individual in the first-come-first-serve range with a different (larger) number (e.g., text/html with deflate encoding). This was noted as a possibility but not considered a significant problem requiring a change to the current setup.

AOB - DNS-over-CoAP (DCoAP)

DNS Update: Clarified that DNS Update functionality is out of scope for the current DNS-over-CoAP document, focusing instead on DNS lookup. A separate, short document for DNS Update might follow if deemed necessary after initial deployment experience.
DNS Push Notifications: Confirmed that DNS Push notifications, potentially mapped to CoAP Observe, are considered within the scope of the current document.

Concise Resource Identifiers (CoRE-HREF)

Carsten Bormann provided an update on core-href, focusing on recent changes in draft-ietf-core-href-17 and open issues:

Test Vectors: One actionable issue on test vectors has been completed. Implementers are encouraged to review existing test vectors.
Issue 93 (Composability): This issue regarding the composability of CRIs was deemed complex and not strictly necessary for the current document's completion. It was decided to defer this for future exploration through hackathons and practical experience, potentially leading to a wiki or a separate document on implementation considerations.
ACE Integration: Discussed the integration of CRIs into the ACE authorization information format (AFCI), adding a new Section 8.3 to explain how CRIs can be used for local parts of URIs, including registering a new object identifier for CRI local parts.
- Editorial Comment: Marco indicated that the text referring to "unsigned integer" in this context might be more accurately replaced with "REST Method Set" from RFC 9237.
Stand-in Tags/Items: Clarified that the document describes the structure of CRIs after stand-ins have been resolved. The document leaves it to applications to define the set of stand-in tags they are interested in, acknowledging the extensibility challenges.
- Editorial Comment: Marco suggested simplifying a "run-on sentence" regarding "evaluate to the original item" for improved readability.
IDNA (Internationalized Domain Names in Applications): Acknowledged the complexity of IDNA in the context of DNS and its relevance only when schemes use DNS. The document advises implementations with scheme-specific knowledge (e.g., HTTP, CoAP using DNS for authority) to use the 2s-key procedure (from the obsoleted RFC 3490) to convert Unicode text in reg-name to ASCII, but cautions against this conversion if the next step is conversion to an II (Internationalized Identifier).
Zone Identifiers (RFC 6874/6874bis): Discussed the ongoing flux around Zone Identifiers (RFC 6874, 6874bis). The document maintains a place for a Zone Identifier in the CRI structure. It acknowledges that conversion to URIs might have limited clarity due to the ambiguities in RFC 6874, which many consider defective. A new test vector was added to clarify the intended interpretation. The text implicitly warns about using Zone Identifiers in URIs.
Determinism Goals: Clarified that CRIs were designed with the objective that two different absolute CRIs (after their own normalization) should not create the same URI. This is a design objective, not a strict promise, due to multiple layers of normalization (CRI-specific, general URI syntax, scheme-specific). This objective does not extend to CRI references, as their resolution depends on the base URI.
New Media Types: Carsten Bormann noted that he updates the table in the appendix with new media types when publishing document versions, leveraging a perfect hashing approach for now, but acknowledging this cannot continue indefinitely. The WG will need to determine a future maintenance strategy for this consistency.

Decisions and Action Items

CoRE Content-Format-based URI Scheme for CoAP (CoRE-CFF-Update)

Decision: Issue 4 (IAB BIS WG notification) to be closed.
Decision: For Issue 22 (Media Type Parameter Quoting), text will be added to the draft to specify a preferred canonical form for recording new registry entries (unquoted if possible, semicolon followed by exactly one space). Existing entries will remain as they are.
- Action Item: Esco to draft the specific text for this change.
Decision: For Issue 23 (Equivalent Media Type Strings), an example will be added to the document to illustrate variations and advise designated experts to check parameter definitions for case sensitivity.
- Action Item: Esco to add the example or delegate to Thomas.
Decision: For Issue 24 (New Registry Column and Provisional Types), a new column for the base media type will be added to the registry, with IANA responsible for filling it and providing clickable links to IANA assignments. The concept of provisional/temporary assignments will be formalized for the 0-255 and ITF ranges.
Decision: For Issue 26 (Lightweight Checks and Expert Review), the concept of "lightweight checks" will be removed. The document will explicitly state that expert review is required for all ranges, including ITF review and ISG approval, drawing on text from RFC 9668 for guidance.
Decision: Issue 28 (Formalizing Temporary Note) is addressed by the changes for Issue 24.
Action Item: The document's status in the data tracker will be changed from "Working Group Last Call" to "Working Group document with issues raised to be addressed".
Action Item: Esco to prepare a Pull Request (PR) addressing the augmented policies for issues 26 and 28.
Action Item: Marco has self-appointed as Shepherd for this document.

AOB - DNS-over-CoAP (DCoAP)

Decision: DNS Update functionality is out of scope for the current DNS-over-CoAP document.
Decision: DNS Push notifications, potentially using CoAP Observe, are within the scope of the current document.

Concise Resource Identifiers (CoRE-HREF)

Decision: Issue 93 (Composability) is deferred. It will not be addressed in the current document but will be explored through future hackathons and practical experience.
Decision: Integration with ACE (AFCI) via a new Section 8.3 is approved.
- Action Item: Carsten to review and potentially revise the text in Section 8.3 regarding "unsigned integer" vs. "REST Method Set" based on Marco's comment.
Decision: The approach to Stand-in tags, allowing applications to define the set of interest, is approved.
- Action Item: Carsten to simplify the sentence "is define to evaluate to the original item" for better clarity.
Decision: The IDNA considerations, advising implementations with scheme-specific knowledge to use the 2s key procedure while cautioning against it before conversion to an II, are approved.
Decision: The handling of Zone Identifiers, maintaining a place in the CRI structure while acknowledging conversion ambiguities, and incorporating PRs 94 and 95 (including a test vector), is approved.
Decision: The disclosure of determinism goals for absolute CRIs (but not CRI references) is approved.

Next Steps

CoRE Content-Format-based URI Scheme for CoAP (CoRE-CFF-Update): A new PR will be created to incorporate the agreed-upon changes. The goal is to ship this document by February.
DNS-over-CoAP (DCoAP): A new version of the document will be released soon, to go in parallel with DNS LPN.
Concise Resource Identifiers (CoRE-HREF): Carsten Bormann will publish draft-ietf-core-href-18 by the end of the week, incorporating the discussed changes and editorial fixes. Following this, the Working Group plans to initiate a second Working Group Last Call, likely for a duration of two weeks.