W3C’s Immersive Web charter needs one disposition record

Summary

On 10 August, the W3C Team opened refinement of a draft Immersive Web Working Group charter. It directed early public discussion to strategy issue 564 and said Members could discuss the proposal on the member-confidential w3c-ac-forum list.
W3C Process section 4.2 requires all issues filed against the charter draft to be formally addressed and their resolutions tracked in a disposition of comments, highlighting issues not resolved by consensus.
The public thread already demonstrates a useful issue-to-text chain: a security reviewer identified a blocking wording problem, a responder linked pull request 862, and the reviewer stated that merger would satisfy the blocker. APA questions produced answers and pull request 865. Both pull requests were still open at the 30 August cutoff.
Confidentiality is not a defect to route around. Process sections 7.2 and 7.3 protect Member-only information and allow authorized, un-attributed public versions that reasonably communicate what must cross a confidentiality boundary.
A public-safe disposition record should identify issue class, affected charter version and section, response, public diff or no-change rationale, consensus state, next authority step and supersession history. It should disclose neither names nor confidential wording, submission counts or identifying commercial detail.
Public comments are evidence and expertise, not an Advisory Committee ballot or a mandate from “the public.” Refinement consensus, Team Decisions, later AC Review and final W3C Decisions remain distinct authority states.

One proposal, two feedback doors

The refinement notice is unusually clear about where discussion may happen. Anyone can follow and contribute to the public strategy issue. W3C Members are also told that they may discuss the same draft through a member-confidential list. The notice does not present one route as ceremonial and the other as real. It points both at one charter-refinement process, under one named Chartering Facilitator, with an estimated endpoint around 7 September.

There are good reasons to keep the doors different. A public repository allows outside specialists, implementers and horizontal-review groups to inspect the proposal, cite a precise line and preserve a durable exchange. A Member channel can carry information whose author is not entitled or willing to publish it: deployment constraints, contractual sensitivities, candid assessments or concerns whose attribution would distort later commercial relationships.

The existence of a protected channel is therefore not evidence of secrecy in the pejorative sense. Nor does the public route become inferior merely because a protected route exists. The governance problem appears one step later. If the two streams can alter the same charter, how does a reader reconstruct the state of the issues without gaining access to confidential correspondence?

The answer is to join disposition, not content. Raw comments can stay where their access rules put them. The public record can still say that an issue class affected a particular section, received a particular institutional response, caused a public text change or received a no-change rationale, and remains open or closed under a named authority state.

The public thread shows what a receipt can do

Strategy issue 564 is not a finished disposition report. It is nevertheless a practical demonstration of the chain that a later report should preserve.

A security reviewer pointed to a small but substantive wording difference. The draft said each specification should contain “sections” covering security and privacy implications, while the current charter template said “separate sections.” The reviewer labelled the comment blocking and explained the closure condition. A responder acknowledged the mismatch, opened pull request 862 and linked a further group discussion about a possible shared security-and-privacy document. The reviewer then said the pull request addressed the blocker and that merger would resolve it.

That exchange has four separately observable states: issue, response, proposed diff and closure condition. At the cutoff, the pull request remained open and unmerged. It would therefore be inaccurate to call the blocker fully closed merely because a patch existed. It would be equally inaccurate to say the comment was ignored.

The accessibility review produced a different shape. APA asked whether plane and mesh detection overlapped, how the WebXR/WebGPU timetable related to implementation and WebGL, why CSS Spatial Layout and haptics appeared in coordination text rather than as deliverables, and what the intended scope was. Responses explained the intended boundaries and linked pull request 865 to revised descriptions. APA said the replies and pull request answered those questions, then raised a further issue about inconsistent use of “detection,” “tracking” and “sensing.”

Here the state is not a binary “APA approved” or “APA objected.” Several questions received answers; one public patch remained unmerged; another naming question continued. A reliable disposition record needs granularity finer than the label on the overall issue.

The examples also show why comment counts are poor governance evidence. Fourteen displayed comments do not mean fourteen independent stakeholders, fourteen issues or fourteen votes. A single comment can contain five questions. Several people can refine one point. Silence can mean agreement, abstention, lack of awareness or no standing to decide. The record should count states, not manufacture representation from activity.

The Process asks for disposition, not a popularity score

W3C’s Process draws a precise sequence. The Team initiates charter refinement by publishing the draft, explaining how to participate, stating an expected duration of at least 28 days and naming a facilitator. During refinement, wide review is completed. The Chartering Facilitator seeks consensus among those participating and makes decisions reflecting that consensus. Where consensus cannot be found, the facilitator may ask the Team for a Team Decision, whose rationale must be documented.

The most important sentence for this case comes next: all issues filed against the charter draft must be formally addressed, and their resolutions tracked in a disposition of comments that highlights issues not resolved by consensus.

That requirement is stronger than “a discussion happened.” It ties input to a response and a resolution state. It is also narrower than “publish everything.” The Process does not erase its own confidentiality levels when it requires an issue disposition.

Before the announced refinement period ends, the Team must choose among initiating Advisory Committee Review, abandoning the proposal or extending refinement. The decision must be announced with the required visibility; a rationale is required when the Team is not moving to AC Review. Those later states should close the refinement record without being backdated into it.

AC Review is a different instrument. Each Member organization may submit one review through its AC representative. Dissent in that phase must be expressed as a Formal Objection. A GitHub commenter during refinement is not casting that review, and a Member posting confidential feedback during refinement is not necessarily filing a later Formal Objection. Treating every expression as the same kind of vote would destroy the distinctions the Process establishes.

Confidentiality can survive a public state index

W3C Process recognizes public, Member-only and Team-only information. Authorized recipients of restricted information must preserve its confidentiality and may not release it to the public or beyond the designated access level. A governance proposal that demanded copied emails, named authors or revealing paraphrases would be defective on its own terms.

But confidentiality need not make the decision path invisible. Section 7.3 says that, for processes with a significant public component, information affecting decision-making may need to be publicly available. Only the Team or an authorized party may change confidentiality. When an author has not supplied a version suitable for a different level, the Team may make available an un-attributed version that reasonably communicates what is required while respecting the original restriction.

That rule suggests a workable minimum. A public record can contain an item such as “Member-only issue class M-03: scope boundary for a proposed deliverable,” identify the affected draft section, state whether the response produced a public edit, and record the Process disposition. It need not say which Member raised it, how many Members discussed it, what product motivated it or reproduce a sentence from the confidential list.

Even an issue-class description may sometimes carry identifying information. In that case, the public entry should retreat to the minimum Process state that can safely be disclosed. It should not disguise disclosure as aggregation, and it should never infer a hidden issue merely to make the table look complete.

This article does not know whether any Member-only comments about this charter exist. The announced availability of a channel is not proof that it was used. The proposed index works in either case: it can record zero public-safe confidential-channel issue classes only when the authorized record establishes that fact, or simply omit any invented count and preserve the channel boundary.

Participation supplies evidence; it does not create a floating mandate

The public route matters because legitimate views can originate outside a Working Group. The Process tells Chairs to ensure that groups consider legitimate views and objections expressed by active participants or by others, including another organization or the general public. That instruction makes outside review operationally relevant.

It does not make every commenter a principal. Expertise about accessibility, privacy, security, internationalization or architecture can reveal a defect. An implementer can identify a schedule that cannot be met. A Member can expose a dependency. Those contributions carry weight through their reasons and evidence, not because the word “stakeholder” grants an executable mandate.

Authority remains allocated. During charter refinement, the Chartering Facilitator seeks and assesses consensus among participants in that process. The Team can make specified decisions and determines whether the proposal advances, is extended or is abandoned. The Advisory Committee later reviews the charter through one review per Member. A final W3C Decision follows that institutional phase.

A good disposition record makes this sequence visible. It says who supplied an input only when disclosure is appropriate, but always says which role supplied the response and which authority closed the state. That prevents two opposite errors: treating public review as theatre because it lacks formal voting power, and treating participation as if it were a universal public mandate.

What one dual-channel disposition record should contain

The common record should remain thinner than either discussion channel. It is an index of governance state, not a second debate archive.

For each issue or safe issue class, it should record:

a stable identifier that survives title changes;
the source class—public or Member-only—without identifying or counting restricted contributors;
a public-safe statement of the issue, if one can be disclosed;
the exact charter section and frozen draft commit reviewed;
the responding role and dated response;
the public pull request, diff or replacement text, or a no-change rationale;
a disposition such as accepted, accepted in part, no change, superseded, withdrawn, deferred or unresolved;
whether the issue was resolved by consensus, or the Process-required indication that it was not;
the next authority step: further refinement, a Team Decision, AC Review, extension or abandonment;
closure time, superseding event and correction history.

The source class is not a weight. “Member-only” should not mean more authoritative than a technically decisive public review, and “public” should not mean more representative than a Member concern. The record describes custody and access, while the rationale explains the decision.

The frozen draft reference is essential. Pull request 862 can satisfy a reviewer's condition only if its text reaches the charter version that advances. A response linked to an open patch is not the same state as a merged patch; a merged patch is not the same as an approved charter. Version identity stops those steps from collapsing.

The limits of the finding

At the evidence cutoff, issue 564 and both cited pull requests were open. The draft still contained placeholders and was still being refined. The active Immersive Web Working Group charter remained the existing charter, shown by W3C as running through 25 September.

None of that proves delay, noncompliance or technical weakness. Refinement is the phase in which comments, drafts and open patches are expected. The 7 September date was approximate, and the Process permits an announced extension. The public state can change after this article’s cutoff.

The directory link to privacy is contextual. It reflects the article’s confidentiality boundary; it does not identify W3C or the Immersive Web Working Group and does not imply endorsement by any directory subject.

The recommendation is deliberately modest: retain the two feedback routes, protect their access rules and publish enough disposition state to make one charter’s lineage inspectable.

W3C Opened Two Doors Into One Immersive Web Charter. It Needs One Disposition Record

Summary

One proposal, two feedback doors

The public thread shows what a receipt can do

The Process asks for disposition, not a popularity score

Confidentiality can survive a public state index

Participation supplies evidence; it does not create a floating mandate

What one dual-channel disposition record should contain

The limits of the finding

Sources

Deeper Profile Context

Strategic Circle

Leadership Alliance

Strategic Circle Briefing

Leadership Alliance Briefing