Summary

  • W3C Strategy issue 553 records APA’s request to opt out of duplicate charter-review notifications. APA said it already checks central charter issues carrying the Horizontal review requested label and otherwise has two places in which to respond.
  • A public comment says applying that label currently creates five linked issues, one for each horizontal-review area. It also warns that a missing accessibility copy could be mistaken for bot failure unless the central record explains the opt-out.
  • Privacy’s workflow differs: a participant said Privacy uses the issues opened in its request repository. The thread therefore describes a federation of review groups, not one back office with one preferred interface.
  • The Math Working Group charter supplies a concrete trace. A bot created accessibility issue 158 from central Strategy issue 545; APA discussed the charter, recorded no concerns, closed the local issue and posted on the central thread.
  • None of this means APA wants fewer reviews, that W3C has accepted the request, or that a ticket is an approval. Issue 553 remained open at the 2 September evidence cutoff.
  • The minimum governance object is a per-group subscription receipt on the central charter issue: requested area, chosen work surface, delivery acknowledgement, authoritative response location, state and disposition.

The word “opt out” hides the real request

The title of issue 553 can be read too quickly. “APA to opt out of charter review notification” sounds, at first glance, as though the Accessible Platform Architectures Working Group wants to leave part of W3C’s horizontal-review system. The body says something narrower.

APA’s concern is duplication. Charter review requests are copied into the horizontal groups’ request repositories. APA says it already monitors the Strategy repository for charter items with the relevant label. A copied issue therefore does not give the group a new charter or a new obligation. It gives responders a second thread in which they may need to state the same result.

That distinction matters because a notification route is not a mandate. APA’s charter assigns the group an accessibility horizontal-review function. Removing one machine-created work item would not remove that function, any concern the group raises, or the later charter process. It would change the group’s inbox.

The public record does not show that W3C has accepted the change. Issue 553 was still open at the cutoff, with no assignee, milestone, label or linked implementation visible on the issue page. Those are facts about the public issue, not proof that nobody is considering it elsewhere.

One Math charter, two issue locations

The request was not posed as an abstract software preference. It followed the review of the Math Working Group charter.

On the central side, W3C Strategy issue 545 carried the draft charter and the Horizontal review requested label. By the cutoff, that issue also displayed completed-review labels for accessibility, internationalization, privacy and security. On the accessibility side, bot-created a11y-request issue 158 said it existed because the label had been added to issue 545 and linked back to the source.

APA’s 15 April meeting minutes show the substantive step: the group considered the Math charter and heard no objections from its accessibility perspective. On 20 April, APA closed issue 158, linked the minutes, stated that it had no concerns and pointed to the response posted on the main thread. The next day, issue 553 was opened from that experience.

The trace shows why a mirror can be useful and why it can be wasteful. Issue 158 arrived inside APA’s familiar repository, joined its project board and made the request visible beside other accessibility work. Yet the conclusion still had to be legible on issue 545, because that is where the charter’s other horizontal-review states and later process labels meet.

Nothing in the trace proves that the duplicate caused a delay or contradictory answer. It does prove that two issue objects represented one accessibility review request. Good governance should describe that relationship explicitly instead of asking readers to infer it from links and repeated comments.

Five copies do not create five equal workflows

A commenter on issue 553 explained that applying the central label currently creates five issues linking back to the Strategy item. Symmetry has an obvious operational appeal. A bot can deliver the same shaped object to each review area, and a missing object can serve as a warning that something failed.

But the thread also exposes the limit of symmetry. Privacy uses the request-repository copies. APA would prefer not to. A Privacy participant said the group could adjust if supplied with the right central query, and a reply provided both a query for open privacy charter reviews and a horizontal-review board. That exchange did not settle a common workflow. It demonstrated that the useful interface depends on the group doing the work.

Uniform evidence is necessary. Uniform inboxes are not.

Horizontal groups have different queues, review rhythms and histories. APA’s own public material separates request intake, issue drafting and longitudinal results across repositories. Its charter lists both the request repository and Strategy issues for new and renewing group charters among the places where engagement is tracked. A central mandate that every group use the same mirror may therefore add coordination work without improving the evidence available to the charter process.

The opposite design is also unsafe. If W3C simply suppresses APA’s generated copy, a person looking at the central issue may see four linked tickets where five used to appear and conclude that accessibility notification failed. The commenter on issue 553 identified exactly that ambiguity. Local autonomy without a central acknowledgement turns intended absence into an error-shaped signal.

Separate the work surface from the authority surface

The repair is a small subscription record attached to the central charter issue, not another universal workflow.

For every requested horizontal-review area, the central record should show:

  1. the immutable source charter issue and draft reference;
  2. the review area requested and the time of request;
  3. the group’s declared work surface—local request issue, central query, dashboard or another public route;
  4. a delivery acknowledgement, including an explicit mirror not requested state where applicable;
  5. the one location in which the group’s authoritative response will be recorded;
  6. the current state, such as received, scheduled, questions raised, no concerns, concerns outstanding or complete; and
  7. the final disposition and any superseding review.

This table should be generated from the same event that drives the bot, so the visible explanation and the delivery action cannot silently diverge. If APA changes its preference later, the record should show when the subscription changed and preserve earlier charter traces. If a group uses a local ticket, the central row should link it. If it does not, the row should affirm that the missing mirror is deliberate.

The authoritative-response field is more important than the notification field. A local ticket may be the best workspace for reviewers. The central Strategy issue remains the place where charter participants need to see whether accessibility, privacy, security, internationalization and architectural review have reached a usable state. Cross-posting every sentence is unnecessary; linking one signed-off response and its disposition is enough.

Labels and tickets are evidence, not decisions

W3C’s general document-review guide helps explain the surrounding machinery. It distinguishes automatic notifications, tracker labels, formal review requests and the later resolution of issues. It also notes that horizontal-review groups can be resource constrained. The guide mainly addresses specification review, so it should not be converted into a charter rule that the sources do not state. Its useful lesson here is narrower: the notification mechanism and the substantive review state are different objects.

A Horizontal review requested label does not prove that a group has received enough information, scheduled work or completed review. A generated issue does not prove a human has read it. A “completed” label records a process state but cannot by itself show the reasoning behind a concern or a no-concerns conclusion. And APA’s “no concerns” on the Math charter did not approve that charter on behalf of W3C.

Authority remains layered. Review groups identify concerns within their fields. Charter proponents respond. A Chartering Facilitator and the W3C Team have defined roles in moving the proposal. The Advisory Committee later uses its own review mechanism. Automation connects evidence among those actors; it does not inherit their powers.

That is also the relevant boundary in Lu Heng’s governance critique. Participation is valuable evidence without becoming a free-floating mandate. A coordination layer can preserve a shared state without dictating every local choice. Here, the bot should make review legible, not turn its preferred repository layout into institutional authority.

What the open issue does not establish

Issue 553 is a useful design record, not a postmortem. The checked sources do not show a missed accessibility review, a contradictory disposition, a failed bot run or a delayed charter. They do not show that all five review groups want to abandon local tickets. Privacy’s comment points in the other direction.

Nor does the absence of a public pull request prove inactivity. The proposal may change, be declined, be implemented elsewhere or remain open. Any later article should test the actual automation output and central status record rather than treating an issue title as a completed decision.

The finding is therefore modest. APA has identified unnecessary duplication in its own workflow. Privacy has explained why a similar copy remains useful in its workflow. W3C can respect both positions if it makes subscription choice explicit and keeps one complete authority-facing review ledger.

Sources

  1. W3C Strategy issue 553 — APA to opt out of charter review notification
  2. W3C Strategy issue 545 — Math Working Group charter
  3. w3c/a11y-request issue 158 — Math Group Charter
  4. APA weekly teleconference minutes, 15 April 2026
  5. W3C — Document Review
  6. APA — Horizontal Review
  7. APA Working Group charter, June 2025
  8. Lu Heng — The Multi-Stakeholder Mirage
  9. Lu Heng — Minimum Initial Specification, Localized Future Decision, and Voluntary Adoption