Summary

  • W3C opened a public Strategy Funnel issue on 26 August 2026 for the Audio Working Group recharter. It describes adoption of Web Speech API as a contemplated substantive change, but says refinement has no expected end date and is not ready for horizontal review.
  • A Working Group discussion proposes keeping recognition and synthesis in one specification, publishing a First Public Working Draft in early 2027 and aiming for Candidate Recommendation in 2028. Supportive comments are proposals, not a recorded group decision.
  • The public charter draft still says speech recognition and speech synthesis are outside the group's direct scope. Its normative deliverables remain Web Audio API 1.1 and Web MIDI API.
  • Web Speech is actively edited and dated 10 August 2026, yet its own status remains Draft Community Group Report. Tests, repository activity and implementation do not silently turn it into a W3C Working Draft.
  • W3C should publish a scope-transfer docket joining the exact source document, revised charter text, review dispositions, W3C decision, Working Group adoption, FPWD and patent state. Each step should change only when the competent actor acts.

One proposal, two public records

The proposed move is easy to describe and easy to overstate. On 26 August, W3C opened Strategy Funnel issue 570 for a successor charter to the Audio Working Group. The issue marks “Considering adopting Web Speech API” as a substantive change in scope. It also records two brakes: the draft is still under Working Group review, and it is not ready for horizontal review. The expected end of refinement is unknown.

A parallel repository discussion adds detail. The current Audio charter expires on 8 November 2026. On 27 August, a participant proposed moving Web Speech into the group as one unified specification covering recognition and synthesis, retaining named editors and seeking FPWD in early 2027 followed by Candidate Recommendation in 2028. A 30 August reply supported the scope addition while acknowledging that milestone estimates are rarely precise.

Those posts are useful evidence of intent. They show a concrete institutional home, a proposed document shape and a rough timetable. They do not show that the Working Group has decided, that the Advisory Committee has reviewed a charter or that W3C has approved one.

The charter text currently points the other way. Its scope covers PCM audio processing, audio-device access, audio capability and cache work, and MIDI-device access. Under “Out of Scope,” it says specific speech-synthesis and speech-recognition functionality will not be addressed directly by the group. Its normative list contains Web Audio API 1.1 and Web MIDI API. Web Speech is absent. A visible drafting note says the deliverables section still needs regeneration, and the completion fields for the listed specifications retain year placeholders.

That discrepancy should not be narrated as institutional failure. It is exactly what refinement can repair. But it must be visible while it exists. A strategy issue announces the proposed change; a charter defines the authority being proposed. Until the same change appears in both, a reviewer cannot know which version represents the request that will eventually go to formal review.

Active technical work is not a maturity label

Web Speech API is not a dormant memorandum. Its editor's report is dated 10 August 2026. It defines browser-facing interfaces for speech recognition and synthesis, identifies an editor, links open issue tracking and points to continuing web-platform tests. The text includes current questions about local processing, language availability, permission and privacy.

Its status heading is equally important: Draft Community Group Report. The Speech API Community Group page says the original group closed on 27 March 2023 and reminds readers that Community Groups are community-run and do not necessarily represent the views of W3C Membership or staff. A separate W3C page records commitments made under the Community Final Specification Agreement.

These records are not contradictions. A technical document can remain active after the original group closes. Editors and contributors can improve it; browsers can expose related functionality; tests can become more complete. None of those events is a substitute for the institutional act of placing the document inside an approved Working Group scope.

The opposite error is also tempting. Draft Community Group Report does not mean valueless, abandoned or technically immature in every respect. It identifies the authority and commitment regime under which the public text currently sits. The correct report preserves both halves: live engineering activity and a still-limited standards status.

Participation starts the change; it does not finish it

W3C's Process separates the stages. A charter must state the group's scope, duration, deliverables, success criteria and expected milestones where available. A substantive recharter requires an Advisory Committee review followed by a W3C decision. The review ordinarily lasts at least 28 days and can be extended to at least 60 days on request. The Call for Review must highlight important changes, explain their rationale and include the disposition of refinement comments.

That path gives issue discussion a proper role without confusing it with approval. Participants surface the proposal and its trade-offs. Chairs and a Charter Facilitator turn it into inspectable text. Horizontal groups can review accessibility, internationalization, privacy, security and architecture. Advisory Committee representatives assess the final proposal. W3C decides whether to confer the authority.

If the charter is approved with Web Speech inside it, another distinction remains. Scope is permission to undertake standards work, not endorsement of the current technical report. The Working Group must still decide what text to adopt and when to request publication. A First Public Working Draft is a public basis for work; under the Process it need not represent full Working Group consensus and does not imply endorsement by W3C or its Members. Candidate Recommendation comes later and requires wider review and implementation evidence.

The 2027 and 2028 dates in the issue should therefore travel as proposals attached to named states. They are not promises. If recognition and synthesis are later split, if horizontal review requires redesign, or if the charter takes longer to approve, the milestones should change without rewriting the history of what was originally proposed.

The licensing record also changes by stages

Moving a document from a Community Group history into a Working Group is not only a change of meeting calendar. It changes the institutional context for patent commitments.

The public Web Speech commitment page records Community Final Specification Agreement commitments. A Working Group operating under the 15 May 2025 Patent Policy has defined participation and royalty-free licensing obligations. Publication of an FPWD begins a 150-day exclusion opportunity under that policy, subject to its detailed rules.

One record should not be presented as automatically replacing the other. The Community Group commitment history matters for provenance. The Working Group participation state, charter scope, adopted draft and later exclusion window matter for Recommendation-track custody. The article identifies no patent claim, dispute or planned exclusion. The governance point is simply that the transition should name which instrument applies at each stage.

This is why the exact source document must be fixed before formal review. “Web Speech API” can refer to a living editor's report that changes between dates. Reviewers need to know which snapshot is proposed as the starting point, whether recognition and synthesis remain one deliverable, and how any required adopted-draft or exclusion-draft information is carried into the charter.

Publish a scope-transfer docket

The smallest useful record is a public docket that follows the work rather than a headline that declares it transferred too early.

The first entries should identify the current Draft Community Group Report, its date and immutable snapshot, its Community Group provenance and its existing commitment record. A proposal entry should name the contemplated receiving group, proposer, date, rationale and whether the unit of work is one specification or more.

The charter entry should point to the first exact draft that actually places Web Speech in scope. It should retain the prior exclusion in version history rather than quietly replacing it. Horizontal-review requests, material issues and public dispositions should attach to that version. The Advisory Committee review window, any extension and the W3C decision should be separate events.

If the charter becomes effective, the docket should then record the Working Group's own adoption decision. FPWD publication and its Patent Policy window deserve another entry. Later Working Drafts, Candidate Recommendations and a possible Recommendation should each retain their own technical, test and implementation evidence. Withdrawal, return to incubation, a split specification or a different institutional home must remain valid outcomes.

The docket need not expose Member-confidential comments, private legal analysis or unpublished patent material. It needs only enough information to show what text moved, who had authority at each step, which reviews occurred and which state is current.

Heng Lu's Policy Mirror offers a narrow test here. A public charter should mirror the work a group is authorized to conduct, and a status label should mirror the decision actually reached. Repository maintenance, participation and implementation are valuable evidence; none is title to institutional authority. This application does not import address-resource ownership into standards work, and it gives the IETF side of the contextual IETF-W3C directory record no role in W3C's decision.

Sources

  1. Heng Lu, The Policy Mirror
  2. W3C Strategy issue 570
  3. Proposed 2026 Audio Working Group charter
  4. Audio Working Group rechartering discussion
  5. Current Audio Working Group charter
  6. Web Speech API editor's report
  7. Speech API Community Group record
  8. W3C Process Document
  9. W3C Patent Policy
  10. Web Speech API Community Final Specification Agreement commitments