W3C home > Mailing lists > Public > www-style@w3.org > September 2004

Re: Problem with 'target-new' property and alternative suggestion

From: David Woolley <david@djwhome.demon.co.uk>
Date: Tue, 14 Sep 2004 07:30:50 +0100 (BST)
Message-Id: <200409140630.i8E6Uo501022@djwhome.demon.co.uk>
To: www-style@w3.org

> >
> > In general, that is not true, as browsers will forward the URI to
> > a proxy and only the last proxy in the chain need actually resolve the
> > address.
> 
> Would you be so kind as to provide an example of this for my humble 
> enlightenment? 

It's not uncommon for a big commercial site to have multiple domain
names mapped to the same site, in order to catch different guesses as
to the site name.  If they then link to the canonical name, rather than
redirecting, you will have multiple domain names that are internal 
relative to the site.

Most people with an efficiently configured browser will have a proxy
configured.  Proxies are used by businesses to control access to 
non-intranet sites, but their primary benefit is that they can cache
web pages and therefore improve responsiveness (current authoring
practices tend to frustrate caching for HTML, but it often still works
for images).

If you have caching, the browser always makes the HTTP connection to the
same machine and passes the whole URL forward.

Proxies can form networks, e.g. the proxy in my office uses the ISP's
proxy as a second level proxy, and the ISP's proxy almost certainly uses
more sophisticated caching schemes that attempt to find the page in 
a number of peer proxies.

Only when you get a proxy that connects to the real web server is it
necessary to resolve the domain name and be able to discover that 
two domain name parts refer to the same machine.

(However, resolving to the same machine name doesn't mean that the sites
are related in ownership, as most web hosting services operate multiple
virtual domains on the same machine, relying on the domain name being
passed forward untranslated, to be used as a tag to select the actual
site, and very small businesses are likely to have their main content
in visible subdirectories of a generic web hosting domain name (often
using hacks with frames to try and hide this from GUI browser users).

Also, smaller businesses often try to pretend that a credit card 
payment processing site is part of their site, even though the domain
name is different - this actually compromises SSL authentication, but
is very common.)
Received on Tuesday, 14 September 2004 06:32:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:32 GMT