httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexei Kosut <>
Subject Re: HTTP/1.1 header problem (fwd)
Date Mon, 23 Dec 1996 10:45:55 GMT
On Mon, 23 Dec 1996 wrote:

> That most certainly is my understanding as well. I did try and ask
> this at the last IETF; but the answer was muddled enough to confuse
> me :-) Anyway I've been going throuh old mail on the httpd-list and
> found several references to myself as an 'idiot' cause I predicted that 
> the early 1.1 servers where going to have a configure flag to allow them
> to pretend to be an 1.0 when connected by a 1.0. The flurry of mail
> following that made clear that the group at that stage considered 1.1
> a dialict of 1.0 and NOT a new version; and there seemed clear consensus
> (by those vocal) that a 1.1 was the _only_ appropriate response; and 
> that anyone who would answer (AOL) in those cases with a faked 1.0
> was to be shot and was most certainly standing in the way of 1.1 
> progress. But quite a few people did see a difficult transition process 
> and where urging server vendors not to waver from the right path :-)

Well, it should be pointed out that until a week ago, there were only
a few (very old) clients that could not deal with a HTTP/1.1 response
to a HTTP/1.0 request. There were also a few (also very old) HTTP/1.0
servers that could not deal with a HTTP/1.1 request. Given that either
way had a few incompatibles, but not very many, we decided that we
perferred sending HTTP/1.1 to HTTP/1.0 requests. Odds are, if (for
example) Netscape Navigator did not acknowledge HTTP/1.1 responses
(and they weren't doing it just to be vindictive - if they just had a
very pedantic interpretation of the non-existant HTTP/1.0 spec), we
would have indeed written Apache to send HTTP/1.0 responses to
HTTP/1.0 requests. But at the time, 99.99% of the browsers worked.

As far as I'm concerned, what AOL has done, since it is easily undone
by them, makes no difference as far as that figure goes, or to what
Apache's response should be.

Alexei Kosut <>      The Apache HTTP Server

View raw message