- From: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Date: Fri, 03 Nov 2017 06:55:03 +0000
- To: Mark Nottingham <mnot@mnot.net>
- cc: Kazuho Oku <kazuhooku@gmail.com>, Willy Tarreau <w@1wt.eu>, HTTP Working Group <ietf-http-wg@w3.org>
-------- In message <576D43E7-C918-40F5-878B-5957599DE668@mnot.net>, Mark Nottingham wri tes: >Again, I'll assert that not every numeric value that could possibly be >transferred in HTTP needs to fit into structured headers; new headers >don't have to use them, and we can add more structures later on if >necessary. I'll add my voice to Mark here: The point is not for this draft to cover every imaginable scenario, but for it to cover as many of the probable ones with as little complexity as necessary. The acid test for me is this: If we had this draft first, and then set about defining all the headers we have to today, would any of them be impossible to express inside the four corners of this draft ? I belive the answer is no. Some of them would look slightly different, but we could transmit the information they contain just fine. The one point where this draft comes up short-ish is timestamps. I would be very easy to convince that we should add a type for timestamps, expressed as a number counting POSIX time_t seconds. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Received on Friday, 3 November 2017 07:16:34 UTC