Summary

  • RFC 3305 described a shift from treating URL and URN as exclusive partitions of URI space to treating schemes as the extensible units; “URL” could remain a useful description, not a separate formal box.
  • The 2002 report was Informational, not a command to implement. RFC 3986 later cited it while recommending “URI” in future specifications—a trace of uptake, not proof of sole cause or universal compliance.

A taxonomy that had stopped helping

In the early Web, the labels looked like a sorting rule. A Uniform Resource Locator pointed to a resource by saying how to reach it; a Uniform Resource Name was supposed to identify it independently of location. A URI, on the classical account, was one or the other. The distinction was intuitive enough to teach, but it became awkward when specifications used “URL scheme,” “URI scheme” and “URN scheme” as if they described different layers.

RFC 3305 was the record of an attempt to clear that muddle. The W3C formed a planning group in October 2000 to consider further work on Uniform Resource Identifiers. It included W3C members and invited IETF experts. Its August 2002 report contrasted the classical partition with what it called a contemporary view: schemes were generally URI schemes, and a scheme could define its own subspaces. http: was a URI scheme; urn: was also a URI scheme, with namespaces beneath it.

Under this view, “URL” did not name a formal compartment into which every scheme had to fit. It remained a useful informal description for a URI whose primary access mechanism helped locate a resource. The shift was subtle but consequential: instead of asking whether a scheme was inherently a locator or a name, one could ask what a particular identifier and its assigning authority made it do. A scheme might support naming, locating or both.

That was a response to documentary friction, not a claim that users everywhere were confused in the same way. RFC 3305 pointed to the mixed language of earlier specifications, including RFC 2396’s generic URI syntax and RFC 2717’s “URL scheme” registration procedures. When neighboring standards name the same kind of extension differently, authors, implementers and registry maintainers inherit translation work before they can even debate the extension’s substance.

Recommendation was not command

The report proposed a joint W3C/IETF model, revisions to RFCs 2717 and 2718 so they would speak of URI schemes, clearer rules for alternative registration trees, and better IANA information about public, obsolete and pending schemes. It also recorded the registry as it appeared in 2002. Its counts—34 schemes in the official list and ten formal URN namespace identifiers—are snapshots from the memo, not current totals.

The document’s status matters. RFC 3305 was Informational: a report to the Internet community, not a Standards Track protocol and not a vote that could compel W3C projects, IETF authors or scheme owners to change their code or vocabulary. It named a coordination problem and offered a path through it. The operational control remained distributed among the specifications that defined schemes, the people who registered them, the IANA-maintained lists, and implementers that parsed or assigned identifiers.

Three years later, RFC 3986 supplied strong documentary evidence that the wording had travelled. Its section on URI, URL and URN says that a URL is a subset of URIs providing a primary access mechanism. It also says a scheme need not be only a name or only a locator, and recommends that future specifications use the general term “URI” rather than the more restrictive terms “URL” and “URN”—explicitly citing RFC 3305. This is direct uptake in a later Standards Track text. It does not show that the earlier memo alone caused the choice, that every standards author followed it, or that the everyday word “URL” disappeared.

The registration story continued on a related but separate track. RFC 7595 later documented Best Current Practice for URI-scheme guidelines and registration procedures. RFC 8820 then returned to scheme authority from another direction: a scheme defines its URI structure, and external standards that prescribe substructure can create problems. Together these later documents show why vocabulary alone could not settle ownership, syntax or registration.

What the record can—and cannot—say

RFC 3305 is best read as a small piece of standards history: a joint planning group identified a terminology collision, proposed language and registry work, and left the change to later specifications and participants. RFC 3986’s explicit citation lets us follow one recommendation into a more formal document. The record does not quantify how much confusion the change removed, establish universal uptake, or assign RFC 3305 sole credit for the modern treatment of URIs.

That restraint is part of the mechanism. A coordination memo can make a common problem legible without possessing the authority to enforce its answer. Publication creates a reference others may adopt; only the later specification, registry procedure and implementation determine what actually changes.