Summary

  • W3C began horizontal review of a proposed successor charter for the Publishing Maintenance Working Group on 17 September 2026. The public strategy record describes charter refinement, not AC Review or approval.
  • The current charter remains the operating authority until 5 February 2027. The successor draft proposes EPUB 3.5-family work, a separate EPUB Annotations track and conditional ISO/PAS work, but its dates are placeholders and several technical outcomes remain unsettled.
  • A charter-handoff receipt should connect the active mandate, immutable draft, each horizontal-review disposition, the later AC Review, the W3C Decision, participant rejoining and the exact deliverables carried forward.

The first signal is easy to misread. W3C strategy issue 573 says horizontal review has been requested for an existing Working Group recharter. It also records an expected end to charter refinement in December 2026. The security review later reported no blocker. None of those facts says that the Advisory Committee has begun reviewing the charter, that the Team has adopted it, or that participating organisations have committed to it.

That boundary matters because W3C’s Process document separates charter work into two formal phases. Refinement comes first. Approval comes later through an Advisory Committee Review and a W3C Decision. AC Review begins with a Team Call for Review and must run for at least 28 days. The Decision can adopt the charter, return it for further work or reject it. Only after adoption does the Team issue a Call for Participation.

The review tickets opened on 17 September are therefore upstream controls. Accessibility, internationalisation, privacy and TAG design-review requests were still open without public comments at the research cutoff. The security request closed after a reviewer found no security blocker. That is a useful disposition for one review domain, not a proxy decision for all reviewers or the Advisory Committee. Public silence in an open tracker also cannot establish that no private concern exists or that no later Formal Objection will be filed.

Two charters, two authority states

The current charter began on 11 February 2025 and ends on 5 February 2027. It is the mandate under which the group is currently maintaining the EPUB family and developing its listed deliverables. Its authority does not disappear because a successor draft exists.

The proposed charter at commit 464cd138 is a different object. Its rendered page is marked Draft. Its start and end fields still contain placeholders linked to a future Call for Participation. Its schedule uses relative T-plus milestones. Those devices allow planning before an approval date is known; they are not promises that work will begin or finish on a particular calendar day.

The substantive change is visible in the current-to-proposed diff. The successor would make EPUB 3.5, EPUB Reading Systems 3.5 and EPUB Accessibility 1.3 explicit future work. Yet the draft itself says that no EPUB 3.5 draft exists and that the final feature list and level of detail remain to be decided. It also identifies EPUB Annotations 1.0 as a separate Recommendation-track deliverable rather than leaving its final packaging unresolved.

This is proposed continuity, not achieved maturity. W3C’s histories show EPUB 3.4 and EPUB Reading Systems 3.4 in Candidate Recommendation work, EPUB Accessibility 1.2 likewise in Candidate Recommendation, and EPUB Annotations 1.0 as a Working Draft at the cutoff. The proposed charter’s requirements for open tests, two independent interoperable implementations and EPUBCheck support are future success conditions. They are not evidence that those conditions have already been met.

The ISO passage requires the same discipline. The draft describes a possible PAS route through the W3C relationship with ISO/IEC JTC 1/SC 34. It makes the applicable specification set depend on timing between current 3.4 work and later 3.5 work, and says the account will need updating when AC voting starts. That is conditional planning. It does not establish that a submission, vote or approval has occurred.

The resource plan is a proposal too. It anticipates at least six active participants, including implementers and people leading editing or tests, about half a working day per week for chairs, editors and test leads, weekly calls, and 0.20 full-time-equivalent W3C staff support. Those figures reveal the operating assumptions behind the scope; they do not demonstrate that the future group has been staffed or that attendance equals consent. The strategy record lists no known concern, but that field is not a substitute for the later AC Review or its objection record.

The missing object is a handoff receipt

A reader now has to reconstruct authority from several surfaces: an active charter, a live draft, an HTML diff, a strategy issue and five review trackers. Each surface is useful, but together they do not yet form one auditable transfer record.

W3C should publish a compact charter-handoff receipt. Its first row should freeze current authority: the active charter URL and dates, present deliverables, and Recommendation-track work inherited from that charter. Its second should identify the exact successor draft by commit hash, record every material scope addition, and mark all placeholder or relative dates as provisional.

A third row should list each horizontal-review request separately, with its status, public response and unresolved action. A closed security ticket must remain one security disposition; it should not silently become an “all reviews passed” badge. A fourth row should begin only when the Team actually calls AC Review, recording the exact draft hash, review window, available comment channels and disposition of Formal Objections.

The final rows should contain the W3C Decision and the participation reset. If the charter is adopted, the receipt should link the decision, enumerate modifications made after review and record the Call for Participation. Existing representatives and Invited Experts then have to be designated or re-designated under the new charter. W3C Process permits a limited meeting-attendance bridge, but participation ultimately means rejoining and accepting the new Patent Policy commitments.

This receipt would not reduce the discretion of reviewers or the Team. It would assign every act to the body that actually performs it. Horizontal reviewers advise on domain risks. The Advisory Committee reviews. The Team decides. Organisations and experts join. The Working Group then operates inside the resulting authority.

That distinction reflects a broader governance rule drawn from Heng Lu’s analysis of the multi-stakeholder mirage: participation is valuable evidence and scrutiny, but it does not become final mandate merely by being visible. Applied here, the point is not that W3C lacks a process. It is that each stage should leave a receipt that prevents one form of participation from being mistaken for another form of authority.

Sources