W3C home > Mailing lists > Public > www-tag@w3.org > February 2003

RE: [Minutes] 24 Feb 2003 TAG teleconf (site metadata, namespaceDocument-8)

From: <Patrick.Stickler@nokia.com>
Date: Thu, 27 Feb 2003 14:45:25 +0200
Message-ID: <A03E60B17132A84F9B4BB5EEDE57957B5FBB43@trebe006.europe.nokia.com>
To: <chris@w3.org>, <www-tag@w3.org>
Cc: <ij@w3.org>



> -----Original Message-----
> From: ext Chris Lilley [mailto:chris@w3.org]
> Sent: 27 February, 2003 13:06
> To: www-tag@w3.org; Stickler Patrick (NMP/Tampere)
> Cc: ij@w3.org
> Subject: Re: [Minutes] 24 Feb 2003 TAG teleconf (site metadata,
> namespaceDocument-8)
> 
> 
> On Thursday, February 27, 2003, 8:55:28 AM, Patrick wrote:
> 
> 
> >>   2.1 Site metadata hook
> >>
> >> [Chris] there is no way to give a URI of a site as opposed to a URI
> >> for a welcome page for it hmm... sites are significant resources,
> >> no? so they should have URIs.....
> >> 
> >>    [Roy]
> >>           /
> 
> PSnc> I would propose that
> 
> PSnc>    http://example.com    denotes the HTTP server
> 
> PSnc> thus
> 
> PSnc>    <http://example.com> a x:WebServer .
> 
> PSnc> and that a separate URI scheme is needed to denote
> PSnc> actual physical machine,
> 
> No, that is not the distinction I was trying to draw. Not between the
> site and a machine, but between a site and a page.
> 
> http://example.com/ is the URI of a page, with a length and content
> and so on. It may also, informally, for humans that can resolve the
> ambiguity and overloading, be used to refer to the entire site.

It should not denote both the site and the page. Rather, if one does
a GET on the site URI, one is (I propose) redirected to a default home page.

> For machines, this is not sufficient. Currently, the concept of 'a
> site' is poorly defined and impossible to cleanly reference, this was
> my point.

Right. So "http://"{AUTH}"/" denotes a site. And just because GET
might return a representation of e.g. "http://"{AUTH}"/index.html"
does not affect the denotation of the site URI.

> PSnc> When one does a GET on either http://example.com or
> PSnc> http://example.com/ we are simply redirected to a default home 
> PSnc> web page,
> 
> 
> Its a welcome page not a home page, 

Err... well, if it makes you happy, we'll call it a welcome
page (rather than the home page of the site ;-)  or let's just
say "web page", eh?

> and it might not be a redirect.

Well, not necessarily using the actual redirection machinery,
but it is a logical redirect, since the server is responding as
if some other URI was specified. I.e., you say http://example.com/
and the server behaves as if you said http://example.com/index.html.

I am proposing that a consistant, standardized
interpretation be introduced (or rather, made official)
whereby "http://"{AUTH} denotes a server, 
"http://"{AUTH}"/" denotes the root site of a server,
and a GET to either of those URIs results in a 
redirection (either actual or logical) to a default
web page for the root site of the server.

Such behavior would be consistent, and also inline with what
most servers already do, and having such an interpretation
gives us distinct URIs to talk about the server, the site,
and any resource in the domain of that site.

And having distinct URIs for those allows us to work with
descriptions of each of them without ambiguity.

Patrick

--
Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com
 
Received on Thursday, 27 February 2003 07:45:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:16 GMT