W3C home > Mailing lists > Public > www-talk@w3.org > November to December 2001

Re: What is at the end of the namespace?

From: Roy T. Fielding <fielding@ebuilt.com>
Date: Fri, 16 Nov 2001 00:57:38 -0800
To: Andy Powell <a.powell@ukoln.ac.uk>
Cc: www-talk@w3.org, uri@w3.org
Message-ID: <20011116005738.C1089@waka.ebuilt.net>
> Phrases like 'network resources' and 'the identified resource is located
> at the server listening for TCP connections' certainly don't convey to me
> the idea that the http URI can be used to denote 'anything that has
> identity'.
> 
> FWIW, I would really like to see your intended meaning clarified (in
> writing) somewhere - perhaps in http://www.w3.org/TR/uri-clarification/ ?

Sorry, but it isn't feasible to include philosophical discussions about
the nature of the Web architecture in a protocol specification that
requires rough consensus on an IETF working group.  The HTTP specification
can only talk about those aspects of the protocol that are relevant to
HTTP.  The URI specification contains much more elaboration on the nature
of any scheme for identification, which is just as applicable to http URLs
as it is to any other identifier.

The scheme name is completely irrelevant to a URI's capacity for
identification -- it merely indicates the syntax for that namespace and,
when used in the context of a user action, some hint to the software
responsible for that action as to how it should go about handling the
identifier.  One does not imply or require the other.  How it is interpreted
depends on the action requested by the user, not on the identifier scheme.
The only thing that matters in setting the scope of what an identifier
can identify is the naming authority, and that space is infinite for HTTP
(along with most other URLs).  The same applies to mailto URLs, but most
people prefer the syntax of http for naming.

In any case, since there is nothing that cannot be identified by an
http URL, including the notion of "nothing" should someone be inclined
to dedicate an identifier to it, I just cannot understand why this
question keeps being raised.  It is an identifier, pure and simple,
and has all the mathematical properties of any other symbolic identifier.
Enumerating those properties in every spec is a hopeless waste of time.

Please note that I am not one of those people who advocate that everything
needs to be identified under one naming scheme -- I honestly don't care
whether people use "http" or "urn" or "mailto" or "extra-special-fred" as
their scheme for identification -- my software (like any good URI software)
is table-based and thus capable of handling any number of schemes once
they have earned the right to be deployed.  My only concern is that the
people creating these new schemes always start out by stating something
which is patently false about the existing ones, and use that false
statement as their justification for creating extra work for others.

....Roy
Received on Friday, 16 November 2001 04:01:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:27 GMT