Summary

  • RFC 9592 retired the Tao of the IETF because one broad, slowly reviewed document could not keep differently changing topics, participation modes and newcomer needs current.
  • Moving guidance into focused webpages and multiple media improves updateability, but currency, authority, archival identity and the exact version a participant observed remain separate evidence questions.

The retired Tao has an unusually clear lineage. RFC 1391 introduced a newcomer guide in 1993. Later editions followed as RFCs. RFC 4677 became the last RFC-form edition in 2006. RFC 6722 then moved the Tao to a webpage in 2012 so it could be updated more easily. Twelve years later, RFC 9592 ended the Tao as a single IETF-maintained document and preserved its last edition in an appendix.

The sequence is not a story of print losing to the web. The Tao had already become a webpage. The failure was more structural: too many topics with unrelated rates of change remained coupled to one review and publication unit.

A monolith makes unrelated clocks wait for one another

Meeting logistics, remote participation, Hackathon practices, Working Group culture, standards terminology and newcomer support do not age at one speed. RFC 9592 says the Tao's breadth and length made review expensive and updates infrequent. Material about the Hackathon could be added and still be buried in an already long document. Its in-person orientation also lagged the rising importance of remote participation.

The publication count makes the friction visible. After the first Tao in January 1993, four more RFC editions appeared over seventeen years. RFC 6722 created a web publication process in August 2012, yet RFC 9592 records only four additional versions over the following eleven years. A mutable medium did not, by itself, produce a modular workflow.

RFC 9592 therefore moved forward with focused guides, shorter documents, videos, podcasts, blog posts, Datatracker improvements and other channels. The current IETF “Getting started” page behaves like a router: it sends a reader toward an introduction, an RFC explainer, Working Group guidance, a new-work guide, meeting guidance, new-participant material, mailing lists and recordings. That architecture lets one subject change without reopening an encyclopedia.

The old webpage contract carried useful receipts

RFC 6722 is worth reading because it did more than say “put the Tao online.” The IESG designated editors. Suggested changes had a discussion venue. An editor assembled proposed revisions. The IESG retained approval. Each version was to show a visible timestamp. Published versions were to be kept at dated archive URLs, with a revision list.

Those controls did not turn the Tao into formal process law. RFC 9592 is explicit that the guide was an informal, community-maintained overview. But the controls gave a reviewer a chance to reconstruct which edition existed, when it appeared and where the replaced edition went.

The successor model has a different strength. Smaller pages can be corrected faster, aimed at a narrower reader and composed into different journeys. RFC 9592 does not define one universal timestamp, archive and approval procedure for every successor page, video, post or channel. That statement must remain narrow: it does not prove that the IETF lacks repositories, backups, editorial review or change histories elsewhere. It says only that retirement of the old contract does not make those receipts automatic across the new surface.

“Official,” “current” and “observed” answer different questions

An official URL answers who is publishing the current representation. A current page answers what that publisher presents now. A dated archive answers what was published at an earlier time. A participant's capture answers what that person actually observed. A formal process document answers what procedure has normative standing. Running practice answers what people and systems did.

These facts can agree, but none substitutes for the others. A new page may correctly explain today's practice while being poor evidence of what an applicant was told last year. An archived edition may prove historical wording without proving that participants followed it. A current sentence may be institutionally approved without creating authority outside the IETF's own scope.

That last boundary matters because the IETF's current introduction repeats the older restraint: it produces voluntary standards and does not control or patrol the Internet. Newcomer guidance can explain an institution, orient participation and improve engineering work. It cannot transform explanatory prose into sovereignty over implementers, operators or networks.

Build a provenance ledger, not another monolith

The remedy is not to reconstruct the Tao under a different name. It is to make modular content independently accountable. For any guidance that affects access, eligibility, escalation, intellectual property, consensus procedure or a consequential engineering decision, preserve a small record: owner, scope, source documents, public URL, publication time, content hash or snapshot, material-change note, review receipt, superseded reference and next review trigger.

The relying team has duties too. A decision record should say which URL and version it used, when it retrieved the page, which formal document controlled if the explanation conflicted, who interpreted it and when revalidation is due. That is especially important when a mutable guide is cited later as though it were an immutable rule.

The evidence chain is then visible: need, owner, source, proposed change, review, publication, version, archive, reader observation, local decision and operational result. Faster publication can shorten the middle of the chain. It cannot remove the ends.

This is where running-code primacy supplies discipline. The guide is valuable when it helps real participants do real work. Its institutional label is not the outcome. Minimum initial specification also points toward smaller common commitments and local accountable decisions. The page should say enough to coordinate; it should not hide who owns the remaining choice. Reality-layer discipline keeps “official page,” “approved process,” “participant belief” and “running result” from becoming one green badge.

Sources