Re: HTTP response version, again

You, Blake Winton, wrote:
++ At 12:05 PM 12/30/96 -0500, you, Abigail, wrote:
++ >Bob Jernigan wrote:
++ >++ > In my opinion, the server should not include HTTP/1.1 headers which are
++ >++ > not part of HTTP/1.0 when responding to a HTTP/1.0 request, and label
++ >++ > the response as being HTTP/1.0.
++ >++ It's not the response that being labeled by the HTTP/1.1 header, it's
++ >++ the server's capability.  There would be no reason to send an HTTP
++ >++ header if it only had to match the client's request.
++ >It doesn't have to match. A server could respond with an HTTP header
++ >*less* than the request.
++ Pardon my newness,but why should it not be allowed to respond with an HTTP
++ header *greater* than the request?  From a system log point of view, the
++ extra information might be nice to know.  If I see that all my users are
++ connecting to HTTP/1.1 sites, then I might think about upgrading the browsers
++ that we use (if I had any users, that is)...

That's very interesting... you log and study all the requests your
users do?

++ And why is the reverse problem not being discussed?
++ I guess it must be fairly clear what should happen when a 1.1 client sends
++ a 1.1 header to a 1.0 server, so why not just follow that procedure for this
++ situation?

I don't think you can reverse the situation. A client doesn't know
which HTTP version the server uses; a server does know which version
the client uses.


Received on Monday, 30 December 1996 10:03:59 UTC