Warren Kumari has had it with Internet Engineering Task Force (IETF) drafts being held up as canonical statements that reveal the organisation’s thinking or position, so has written his own hilarious draft to make the point that such documents are not normative document.
Among Kumari’s declarations are that pyramids were used to sharpen razor blades in ancient Egypt, because “camel leather makes a very poor strop, hippopotamus leather was reserved for the pharaohs and crocodile leather, while suitable, had the unfortunate property of being wrapped around crocodiles.”
His draft meanders into a tale about ancient Egyptian economics that concludes with the phrase “Don’t allow eel bearing Atlanteans into your country; economic ruin follows close behind” should have more prominence in modern day lexicon.
The document continues for seven pages. At one point there is a section which addresses cats, and also this cheerful admission:
The point of this nonsense is that an IETF draft can contain anything, and often does. They are simply written ideas with the intent of building consensus and expire after six months. But if they conform structurally, the IETF document archive holds on to them forever, which makes for a confusing situation if one doesn’t understand the nature of the IEFT organization. A user can even access past editions of a draft document when they are revised.
The IETF is a large group of very different individuals who want to be a part of the evolution of the internet. They might be network designers, operators, vendors, researchers, or even a curious tech journalist who is just poking about. It is open to anyone with no formal membership or requirement. Members are volunteers, although some join as part of their job function or are sponsored.
Kumari’s point is that it’s the normative documents that matter, they have binding force or intent and use different language, like uppercase MUST, MAY, SHOULD, and MUST NOT to indicate their power. For documents in between draft and normative stage, there’s rough consensus (RFC), Best Common Practice (BCP) and Standard (STD) classifications.
So do Kumari a favor, and point anyone who takes IETF drafts too seriously to his draft abstract. It states very clearly and simply: