Session Date/Time: 27 Feb 2024 22:00

RSWG

Summary

The RSWG session primarily focused on the ongoing Working Group Last Call for the RFC 7990 updates. Key discussions revolved around the preferred approach for publishing revisions during the last call, clarifying terminology for RFC formats, the role of various stakeholders (RPC, tools, stream managers) in reviewing the document, and the feasibility of automated regeneration testing for various publication formats. A decision was made to publish a new revision (05) of the draft. Several technical issues were identified for further discussion on the mailing list.

Key Discussion Points

Note Taker: Russ Housley initially offered to take notes but encountered issues with the Etherpad. Robert Sparks assisted with note-taking.
Working Group Last Call (WG Last Call) for RFC 7990 Updates:
- Revision Strategy: Paul Wouters noted that previous rollups (04) incorporated editorial and non-controversial changes. Robert Sparks advocated for frequent revisions (an 05 draft) during the WG Last Call rather than just list rollups, to facilitate review.
  - Decision: A sense of those present indicated no disagreement with publishing draft-ietf-rswg-rfc7990bis-05. Paul Wouters agreed to prepare it, including recent rollups and a pending pull request.
- Open Issues Identified by Authors:
  - New Definitions: Jay Daly's proposal for adding new definitions, specifically to clarify "definitive format" (RFC XML) and "publication formats" (e.g., PDF, plain text, HTML) to resolve confusion about the term "version." Pete Resnick cautioned against using "human readable" as a descriptor. Paul Wouters indicated willingness to add a simple definition in the introduction but preferred not to sprinkle new terms throughout the document unless explicitly agreed. Eric Rescorla sought clarity on consistent usage of such terms.
  - Removal of Prohibition on Comments: Martin Dorst raised concerns about removing the prohibition on comments, arguing that such hidden information (not appearing in publication formats) could be problematic. John Levine questioned the consistency of this argument, noting metadata also doesn't always appear in output. The discussion was deferred to the mailing list.
  - Explicit 9280 Updates: Elliott Noss proposed explicitly calling out changes that impact RFC 9280, especially historical policies in Section 7 (e.g., 7.6 regarding RFCs not changing once published), and potentially a new policy on "consistency of the series" as a reason to update. He suggested using the "old/new" framework from RFC 9280.
- Stakeholder Review and Feedback:
  - Brian Carpenter emphasized the need for input from the tools side and the RPC (RFC Production Center) to ensure implementability.
  - Paul Wouters added stream managers to the list of important reviewers, particularly to avoid late surprises from the RSAB (RFC Series Advisory Board).
  - Eric Rescorla clarified that the RSAB is the proper venue for rejections, but chairs should proactively ping stakeholders for review as the document stabilizes.
- Potential Second WG Last Call: Paul Wouters suggested a possible short second WG Last Call, focused on objections, especially if significant changes are incorporated before the current one concludes. Chairs acknowledged the point, emphasizing ensuring reviewers are aware when the document becomes stable.
- Archived Versions Terminology: Discussion around distinguishing the current "definitive version" from previous, superseded versions (Martin Dorst's "archived" point). It was acknowledged that precise vocabulary is important for legal and historical clarity (e.g., "this is the thing" vs. "things that were before the thing"). Robert Sparks highlighted the cost of maintaining archives and advocated for version control systems, not just date-stamping, to manage document history effectively. Alexis outlined the importance of a persistent URL for the current definitive version, but argued against including specific implementation details of archiving in this document.
- Appendix Advice: Regeneration Testing: Jay Daly questioned the advice in the appendix regarding testing whether document regeneration produces the "same result" across different output formats (plain text, HTML, PDF). He argued this is challenging and potentially unachievable for PDF and HTML, citing issues with PDF diff tools and expected HTML changes. Eric Rescorla defended the need for automated verification, especially for text and HTML, to ensure tool chain correctness. Jean-François C. Morin provided examples of unhelpful PDF diffs. Elliott Noss stressed the need for the RPC's interpretation of these policies. Paul Wouters suggested focusing verification efforts on HTML, assuming PDF conversion from verified HTML would likely be sufficient, while Alexis indicated this sets too high a bar without demonstrable tools. Robert Sparks agreed on the value of CI but questioned making it a requirement in the policy document. Eric reiterated its policy nature and the value of automated verification over none.
"Not ASCII" Document: Briefly mentioned as the next agenda item after the 7990 updates, but no discussion took place.
RSAB Community Consultation: Elliott Noss channeled Mirja K. and suggested the RSWG provide input to the RSAB on how the RSAB should consult the community once this document reaches consensus.

Decisions and Action Items

Decision: The working group agreed to publish a new revision of draft-ietf-rswg-rfc7990bis-05, incorporating recent rollups and a pending pull request. Paul Wouters will prepare this.
Action Item: Chairs (Russ Housley) will explicitly reach out to representatives from the RPC, tools side, and stream managers to ensure they review the document and provide feedback during the WG Last Call.
Action Item: Paul Wouters will draft proposed text for clarifying "definitive format" and "publication formats" for discussion, likely as a pull request or list comment.
Action Item: Discussions on the "prohibition on comments" and the specifics of "regeneration testing advice" (especially for HTML/PDF) will continue on the mailing list. Specific text proposals are encouraged.
Action Item: Elliott Noss will start a mailing list thread to gather RSWG input for the RSAB on how to best consult the community regarding the approval of this document.

Next Steps

Paul Wouters to publish draft-ietf-rswg-rfc7990bis-05.
Continue active discussion on the mailing list for open issues, particularly:
- Refined terminology for RFC formats and archived versions.
- The removal of prohibition on comments.
- Specific wording for regeneration testing advice, considering technical feasibility for different formats.
- Elliott Noss's proposal for explicit 9280 updates.
Chairs to ping relevant stakeholders (RPC, tools, stream managers) for review.
Elliott Noss to initiate mailing list discussion regarding RSAB community consultation.
The working group will be notified when the document is considered stable to allow for final review opportunities.