When the DID Resolution spec next publishes, what has happened to DID URL dereferencing?

The specification is real and its dereferencing sections are formally marked at risk by the working group itself. The deliberation below is invented — a simulation of how the question settles, published before it does.

The question

DID Resolution v1.0 is a W3C Candidate Recommendation Snapshot, published 6 August 2026. Five of its features carry an at-risk marking — four on DID URL dereferencing, one on resolution architectures — the working group's own signal that the material is likely to be changed or removed. The specification cannot advance before 3 September 2026, and exit from Candidate Recommendation requires at least two independent interoperable implementations of every feature. What has happened to the dereferencing sections when the specification next publishes?

Resolves 2026-11-03.

How this resolves

Checked on 3 November 2026 by reading the most recent published version of DID Resolution v1.0 at w3.org/TR/did-resolution/ and comparing the status of the DID URL dereferencing sections against the branches above. If no new version has published by that date, the entry records UNRESOLVED AS OF 3 November 2026 and is checked again when the specification next publishes, appended to this page. Final check 3 February 2027; still unpublished then, it closes UNRESOLVED.

The room

Working group participants arguing the value of the feature, implementers reporting what they have built, and the Team, which verifies a transition request and can withhold it. Roles and bodies are real; the individuals and their exchanges are invented, and no real participant is named or depicted.

The branches

• 1
  Dereferencing survives, revised toward the provisional modular approach — the alternative algorithm in §6 displaces the one in §5.

  The run's leading branch, by a clear margin over the rest.

• 2
  Dereferencing survives close to its current form, with §6 remaining informative.

  Second, and well behind the first.

• 3
  Dereferencing is cut from v1.0 — deferred to a separate document or a later version, so the spec advances without it.

  Ranked below the first two.

• 4
  The at-risk marking stands and the sections publish unchanged and still flagged.

  The run's least-favoured branch.

This ranking is the simulation's own, and it is uncalibrated — we have no record of how often rankings like it have been right, and we do not publish one.

The three implementer reports are assumed to converge on §6 favorability, but real-world implementation feedback often reveals edge cases or integration costs that shift working group consensus unexpectedly.

What would change the answer

The working group has said it is likely to converge on a revision to the dereferencing algorithms. If that convergence is minuted before publication, the question is largely settled in public before we would learn anything from it. Five features carry an at-risk marking, four of them on dereferencing and one on resolution architectures. They can move independently: the section could survive while the result structure that serves it does not. Exit from Candidate Recommendation requires at least two independent interoperable implementations of every feature. A feature nobody implements cannot advance, so the implementation reports govern this more directly than any argument in the room does. The specification cannot advance before 3 September 2026. Nothing before that date resolves it.

How this was run

A grounded multi-agent simulation over the actors listed above, across 7 moments. Off-the-shelf models. It is a simulation, not a forecast from a model with a track record.

AI-generated simulation. DID Resolution v1.0 is a real W3C Candidate Recommendation; the participants, discussions and outcomes shown here are invented to demonstrate the method. The outcome is not yet known. Not reporting, not a prediction service (FTC 16 CFR 465).