What is an RFC, and how does an internet standard get made?
A Request for Comments is a numbered document published by the IETF describing a protocol, a procedure or an observation — and the name is a historical artefact that persists, because many RFCs are not requests for anything and are the definitive specification of how the internet works.
Why the name. The first was published in 1969, deliberately informal, to avoid the impression that a small group was issuing decrees. The tentative name stuck, and the tone it set — collaborative rather than authoritative — shaped the culture that followed.
What an RFC actually is. Once published, an RFC is never changed. Corrections are issued as errata; replacements are new RFCs that obsolete the old one. This is why you must check whether the RFC you are reading has been superseded, and why old numbers persist in conversation long after the content moved.
The categories, which people conflate:
Standards Track — Proposed Standard, then Internet Standard. Most things people call standards are Proposed Standards and never advance further, because deployment matters more than the label.
Informational, describing something without standardising it.
Experimental, Best Current Practice, and Historic for obsolete material.
April Fools' RFCs, a genuine tradition, published annually and occasionally mistaken for real specifications.
How a standard is actually made:
An individual or group writes an Internet-Draft, which expires after six months unless renewed — deliberately, to prevent abandoned proposals accumulating.
A working group discusses it on a mailing list, which is public and archived.
Rough consensus and running code is the guiding principle: not unanimity, and not voting, but the absence of sustained unresolved objection — plus, importantly, evidence that it actually works when implemented.
Last call, then review by the steering group.
Publication, with a number.
Why it works. Participation is open to individuals rather than to companies or countries; there are no formal votes and no membership fee; and adoption is voluntary. A standard succeeds because people implement it, which is a very different enforcement model from most standards bodies.