- From: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Date: Sun, 13 Jul 2014 21:31:19 +0000
- To: Erik Nygren <erik@nygren.org>
- cc: Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
In message <CAKC-DJjsSHr+WRQ1r7etoYOBW63vLpCOUY9SvcqnkSNB30FuZw@mail.gmail.com>, Erik Nygren writ es: >Where does the 256 octet minimum come from? That seems like an arbitrary >value. Is it too low? The minimum values have ended up mattering in >other protocols (the IPv4 minimum bleeding over into impacting DNS, etc) >so we should be careful not to set it too low. It's my back-of-the-envelope calculation for minimum interop. 256 bytes is enough to do "GET /" and get a redirect back, and to "GET /robots.txt" and get a "deny all" back. There are plenty of non-browser web-apps which can function in 256 bytes and therefore there is no reason for us to insist the accept larger frames. (I'd appreciate somebody else confirm my estimate) -- 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 Sunday, 13 July 2014 21:31:42 UTC