Re: About that Host: header....

Ok, I believe we have consensus (not even just rough concensus)
on my option 4.  I've been corresponding with John Klensin 
(slowly; John's been travelling), and he also now agrees to this solution.

To remind people:

4) (Host+Error) Suggested at the IETF meeting:

a) Add host header, with improved wording in the specification.
b) Require 1.1 server to accept full URL from 1.1 or later client.
  (so far, same as option 3).
c) Require server to generate an error if a 1.1 client is detected, and no host
information present (or more strongly, at the expense of extra bytes on
the wire, no host header present). 
Transition to requiring full URL in future HTTP version, once 1.0 servers
have been replaced.

The final key observation, which John was not aware of until
yesterday (with the resulting response of "Arggh" from John), 
is that HTTP 1.0 servers always close their connection 
after a request, even an error request.  A retry after an error,
which would be required if the protocol were fixed properly,
would cost another full connection open, not just a extra packet exchange.
This second open, would exacerbate the Internet's congestion problem.

John was mislead by the fact that FTP and telnet are not so stupid to
close the connection after error.  I guess we just have to
regard HTTP 1.0 about as stupid as can be imagined.

I will draft changes implementing the above and circulate for review
in the next day.
				- Jim Gettys

Received on Wednesday, 27 March 1996 08:14:46 UTC