Summary

  • The operative 2025 W3C Process assigns three TAG seats to Team appointment and requires both the Advisory Board and TAG to ratify the choices by two-thirds secret ballots. February 2026 appointment notices for Brian Kardell and Heather Flanagan say both bodies did so.
  • The 25 October 2026 Process draft would retain Team selection and TAG’s two-thirds vote while dropping the Advisory Board’s. The AB said in July that it no longer thought routine ratification belonged to it; the public issue thread treats possible help with a Team–TAG deadlock as a separate question.

Two names, four votes, one proposed deletion

On 6 February, W3C announced Brian Kardell’s appointment to its Technical Architecture Group for the 2026–2028 term. The notice says the Team chose him and both the Advisory Board and TAG ratified the choice. On 12 February, it announced Heather Flanagan with the same two approvals. Two real appointments went through the rule that the next Process draft proposes to shorten.

The operative Process divides TAG membership among eight people elected by the Advisory Committee, three appointed by the W3C Team, and Tim Berners-Lee as a life member. Appointments are meant to fill gaps in technical background and skills left by an election. W3C requires a way for the wider community to propose candidates and specifically solicits input from TAG members, the Advisory Committee and working-group chairs. An affiliation rule can rule a candidate out when an employer already has an elected TAG seat. The February notices say several candidates were disqualified on that basis. They do not publish names or secret-ballot margins.

Under the current rule, the Team’s chosen appointee needs a two-thirds secret ballot from TAG and a separate two-thirds secret ballot from the Advisory Board. A peer group can reject a candidate it considers unsuitable for architectural work. The broader elected advisory body can apply a second institutional check. Neither ballot chooses the original candidate; that pen stays with the Team.

The 25 October draft deletes the Advisory Board ballot from section 3.3.3.4. TAG’s two-thirds approval remains; so do the Team’s selection role, the public nomination route and the affiliation and term constraints. This is a proposal in the next Process, not the rule that governed the February appointments. It changes one lock on the gate, not the ownership of the gate or the three seats behind it.

The referee put down the whistle

The official discussion in issue 810 has a better argument than “less bureaucracy.” In 2023, one contributor defended the Advisory Board’s vote as a guard against an insular TAG. Another asked why a body concerned with organization and process should have the same approval power as the technical group that would work with the appointee. The point of disagreement was concrete: independent distance can catch a closed club, while a second full veto can put technical fit in the hands of a group with a different job.

By 2 July 2026, a participant reporting Advisory Board feedback said the AB did not feel it should be involved in ratifying TAG appointments. W3C’s pull request 1178, merged on 22 July, says the change was made per that AB decision. This is a relinquished power, not a power wrested away over a recorded objection.

The issue thread then separates a second question. If the Team proposes someone and TAG withholds approval, should the Advisory Board help resolve the dispute? A July response says the AB would be happy to advise. Other comments debate whether advice should ever become an arbitration power. The current draft keeps the TAG vote and removes the routine AB vote; it does not write a new binding AB appeal into the appointment section. A deadlock could therefore end in another Team choice, a delay or some other process outside the text reviewed here. The public documents supply no timetable or required escalation for that case.

The February announcements make the stakes more specific. The Team split that year’s two appointments, saying time constraints and an upcoming TAG meeting made it undesirable to keep the group waiting for both names. Under the old rule, both bodies ultimately ratified each announcement. The record does not tell us either ballot was close or that the AB delayed either person. It does show that appointment timing already mattered enough for W3C to announce one name before the other. A future failed ballot would land on a working group with real work to do.

Keep the technical veto, show the choice

The TAG’s own public description makes clear why its members have a say: they steward Web architecture and review technical work from inside and outside W3C. A two-thirds TAG vote is a serious check on Team selection. It is also a peer judgment by people who will share the table with the appointee. The draft can trust that judgment without pretending it answers every question about how a shortlist was built.

A proportionate public account would identify the skill gap, the nomination window, the affiliation screen and the reason a particular expertise was needed. It could disclose that a candidate pool was narrowed without exposing private names or secret ballots. W3C’s February notices already begin this work by explaining the public call, screened affiliations and the split timetable. The next Process can make that account routine. The remaining hard question is the deadlock: who moves next when TAG says no?

Leaving that choice to good manners is a strange economy for an organization that can specify a two-thirds secret ballot to the last decimal.

Sources