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

Re: [metaDataInURI-31]: Initial draft finding for public review/ comme nt.

From: Roy T. Fielding <fielding@apache.org>
Date: Mon, 14 Jul 2003 23:15:34 +0200
Cc: www-tag@w3.org
To: "Williams, Stuart" <skw@hp.com>
Message-Id: <4E6B7BD5-B640-11D7-884B-000393753936@apache.org>

> Let's see if I've got this...
>
> <http://www.vw.com/touareg/> (with no surrounding context - which I 
> guess is
> itself a context) identifies a website about a brand of car.

Sorry, I meant to add "in plain text".  In structured text, I would
assume there is some metadata that defines the context.

> The sentence:
>
>   "I want one of these cars: <http://www.vw.com/touareg/>."
>
> indirectly identifies the particular brand/model by reference to the 
> website
> about the brand.
>
> The sentence surrounding the reference to the web-site establishes 
> that the
> referent is the brand/model and not the website itself.
>
> ie. the referent of a reference made using a URI is generally 
> dependent on
> its surrounding context.

Mostly.  I would say that the referent of the use of a URI has only
indirect relation to the referent of the URI.  It is therefore false
to say that all uses of a URI must be consistent, even if we require
that all referents of the URI be consistent.  The interaction method
must be orthogonal to resource identification. Otherwise, the
constraint on "GET must be safe" wouldn't make any sense.

....Roy
Received on Monday, 14 July 2003 17:15:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:18 GMT