Re: Confused by HTTP 1.0 vs HTTP 1.1

Matthew Stanfield <mattstan@blueyonder.co.uk> writes:

> I've developing some software that uses sockets to communicate with web
> servers using HTTP. I've seen both HTTP/1.0 and HTTP/1.1 used in GET
> headers. EG. "GET / HTTP/1.0" and "GET / HTTP/1.1". So far I've been using
> HTTP 1.0 in my GET headers but while browsing this list's archives I saw
> that HTTP 1.1 is in use (EG. the thread "HTTP 1.1, proxy servers, and
> failed connections") and wondered whether I should in fact be using 1.1
> instead.

Probably.
 
> Having run some tests it seems that some 'mainstream' web servers (EG.
> www.google.com) don't seem to respond to a HTTP 1.1 GETs. While others (EG.
> www.yahoo.com) do and act normally.

Google does use HTTP/1.1 (though in checking that, I discovered that
thier server does not support either OPTIONS or TRACE).  If they are
not responding as you expect, then you are probably not constructing
the request correctly; there are differences.  See RFC 2616:

  ftp://ftp.isi.edu/in-notes/rfc2616.txt

(most likely you are not including a 'Host' header - it was not
required in 1.0, but is in 1.1).
 
> Can someone explain to me whether I should be using HTTP 1.1 at all and, if
> so, under what circumstances?

It has far better control of caching, and properly supports persistent
connections; if either would be useful, then certainly.

-- 
Scott Lawrence        
  Actively seeking work 
  http://world.std.com/~lawrence/
  [ <lawrence@world.std.com> is deprecated ]

Received on Tuesday, 4 March 2003 14:30:06 UTC