Summary

  • Steve Crocker's RFC 1 exposed Host Software work before the ARPANET was operating, while RFC 3 invited timely, incomplete and even unanswered contributions but still required a number, author, affiliation, date, title and distribution path.
  • Publication made a proposal findable and open to reply; it did not make the proposal official or turn it into an Internet Standard. The later RFC system added status, review, streams and archival stability without erasing that distinction.

A document arrived before the network

The date at the top of RFC 1 matters: 7 April 1969. The first Interface Message Processor would not reach UCLA until September. Crocker's Host Software note therefore was not a report on a stable network. It was a working object sent ahead of one, an attempt to make several hosts behave coherently before anyone could demonstrate the whole path.

The note records a division of responsibility. Bolt Beranek and Newman was building the packet-switching equipment; the host groups had to agree on software above it. Representatives of the initial sites had met, and a smaller working group had continued debating how hosts should communicate and what experiments they should attempt. RFC 1 puts requirements, terminology and tentative mechanisms into circulation while acknowledging that parts of the design still needed development.

That is a different kind of evidence from a finished specification. It proves that a named author at a named institution put a particular proposal into the shared record on a particular date. It does not prove that every site accepted it, implemented it or kept it. The document is valuable precisely because readers can see the distance between the first written design and the system that followed.

RFC 3 lowered the threshold without removing the receipt

RFC 3, Documentation Conventions, made the publishing bargain explicit. Membership in the Network Working Group was not closed. A note could come from any site and from anybody. It could contain a philosophical position, a specific technique without a full introduction, or a question with no attempted answer. Crocker set the minimum length at one sentence and encouraged notes to be timely rather than polished.

That radical permission was paired with a small but durable receipt. Every note still carried the series name and serial number, the author and affiliation, the date and a title. Copies went to a named distribution list, and sites could reproduce them locally. Low editorial friction did not mean anonymous speech or memory by hearsay. A contribution could be incomplete because the record made it possible to identify, answer and supersede it.

This is the mechanism behind the title “Request for Comments.” It was not merely a friendly name. It described what the object was for: a contribution waiting for contact from others. Its number made the contribution citable. Its incompleteness made response legitimate. Those two properties had to coexist. Without a stable reference, a reply would not know what it was revising; without acknowledged provisionality, a reply could look like insubordination to an authority that the group did not possess.

Writing could manufacture authority by accident

RFC 3 names the danger directly. People tend to treat a written statement as authoritative merely because it is written, while authors hesitate to circulate something unpolished. Those incentives reinforce each other: the more official a document looks, the more reluctant a junior participant becomes to challenge it; the less challenge it receives, the more settled it appears.

Crocker's later recollection in RFC 2555 supplies the institutional scene. The early group was junior, informal and unchartered. Its members did not know whether an official protocol authority would arrive and take over. Crocker worried that organizing the notes might look like claiming power he did not have. The chosen label emphasized dialogue instead of control.

The response was not to eliminate form. It was to separate form from finality. Number, author and date made an assertion accountable. The word “request” limited the claim that those formal elements could make. This is a subtle governance design: form creates memory, while status constrains how much power that memory may exercise.

Distribution was part of the protocol

An unfinished note is useful only if it reaches the people who can test or contradict it. RFC 3 listed recipients and left reproduction to each site. In Crocker's retrospective, each institution sent copies directly to the others rather than waiting for a central redistribution point. The arrangement reduced delay, while the SRI Network Information Center maintained the central collection.

The topology divided two jobs that are often confused. Peer distribution moved current work quickly. Repository custody made the accumulated record retrievable. A single central dispatcher would have slowed conversation; a purely peer-to-peer exchange without a collection would have made later reconstruction fragile. The early system used both.

That distinction survives in modern engineering even though the medium has changed. Mailing lists, repositories and issue trackers move discussion; a publication series fixes a stable result. If the conversational channel disappears, readers may see a final artefact without the objections that shaped it. If the archive disappears, the conversation cannot establish which version or proposal it was answering.

The first RFC was not the first Internet Standard

RFC numbering later became a mark of extraordinary durability, and that success makes the original boundary easy to forget. A number identifies a document in a series. It does not, by itself, state the document's maturity, consensus or normative force.

RFC 1796 made the distinction explicit in 1995: not all RFCs are standards. Informational, Experimental and Standards Track documents share a publication channel, and their status must accompany a responsible citation. Calling every RFC an Internet Standard removes the field that tells an operator whether a text records an experiment, describes information or carries standards-track weight.

The point is not that early RFCs were unimportant. It is that importance and authority are different variables. RFC 1 mattered because it exposed problems that hosts had to solve together. Its propositions gained or lost force through response, implementation and later documents, not through the serial number alone.

A provisional series became an archive

Crocker wrote in the 30-year retrospective that he had expected the notes to be temporary and the series to disappear after the network was running. The opposite happened. The working record outlived the hardware, the original Host-to-Host protocol and the small group that produced it.

RFC 8700 describes the resulting evolution. Rough exchange now happens chiefly through email, working groups and Internet-Drafts. RFCs are produced through defined streams and review processes, edited for clarity and preserved as a canonical record. The contemporary series supplies a point of stability to implementers who cannot negotiate every protocol decision again.

This change does not refute the 1969 design. It reveals the lifecycle of a record. At the beginning, low friction was needed to discover the design. Later, stronger process was needed so global implementers could rely on a published result. The mistake would be to apply the final publication burden to the first unresolved question—or to apply the first-note informality to a text on which the world will build.

The answer to an RFC was often another artefact

Comments did not have to appear as marginal notes on the original page. A later RFC could revise terminology, reject an approach or introduce a working mechanism. Code could expose an incompatibility that prose had missed. A host coming online could turn a neat proposal into an observed failure. The response path therefore crossed documents and implementations.

For historical and operational readers, that means a single RFC is rarely a complete decision record. One must look for the documents it updates or obsoletes, its status, related drafts, implementation reports and the behaviour that actually shipped. The early number is an address into the conversation, not a substitute for the conversation.

That is Crocker's deeper contribution to institutional legitimacy. A group without a charter did not solve its authority problem by speaking more confidently. It created a format in which uncertainty could be named, attributed and answered. Legitimacy grew through the traceable interaction among proposal, response and running system.

Sources