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

Re: httpRange-14

From: Norman Walsh <Norman.Walsh@Sun.COM>
Date: Fri, 01 Aug 2003 16:28:36 -0400
To: www-tag@w3.org
Message-ID: <87he51unpn.fsf@nwalsh.com>

Hash: SHA1

/ Tim Berners-Lee <timbl@w3.org> was heard to say:
| Sorry, but the semantic web architecture absolutely needs the idea
| of information resources.  The RDF identifier foo#bar is
| used by dereferencing foo and parsing it a  get information.

So, suppose it does that, without giving any thought to what 'foo' is
before issuing a call to the underlying URI resolver and consuming any
RDF that comes back. What goes wrong?

|>   The URI does
|> not change, nor does the resource, so any claim that the scheme
|> causes the resource to fall into one category or another is false.
| The scheme is a scheme for identifying things in a certain class.

So, what do you assume axiomatically about urn:foo? I assert that on
my machine urn:foo resolves without any trouble into an RDF document
that contains some assertions about urn:foo. Will your system do the
same? It better.

| particular range is architectual mistake.  For example, one must be
| able to define  ISBN number scheme and a car license number scheme
| such that their ranges are distinct, so that you can't put an ISBN
| number on a car licence plate for example.

Why must one do that? How is the assertion that my licence plate is an
ISBN inherently different from the assertion that I am a garden gnome?

| All I am saying is that the mapping the URI identifies that which is
| common among the set of representations, and that that is not the
| subject, but the information content. And the web actually depends on
| that invariance to be useful.


I can't get my head around this.

http://norman.walsh.name/ identifies something. It returns any one of
four different representations, depending on how you ask for it.
There's nothing that's true for any of those representations that's
true for all of them. So how can it matter what the underlying "thing"
identified by http://norman.walsh.name/ actually is? You can't see it,
smell it, taste it, touch it, or interact with it in any way except by
the representations it returns. And it isn't any one of those, it's
something else.

| (a)  In the proposed architecture the assumption would be true, not
| false.
|    when you argue the demerits of the architecture, you have to use
|    its rules.
|    So in my architecture, the assumption is that the URI can be used
| to refer to the web page by anyone who has come across the web page.
| (b)  In your alternative architecture, where does one get the explict
| information from - which allows one to refer to a web page by its URI
| without meaning bridge or a person?

Why does the system have to know *anything* about the URI other than
the fact that it gets representations back (sometimes) if it asks for

| It is connected with GET, but not defined by GET.
| A picture of a car is for me a concept itself.

Sure it is. And it's distinct from the concept of the car, I get that.
And I can imagine having URIs for both of them. And I can imagine
either one or both or perhaps neither returning representations.

[I have to run. I'll read the rest of the message this weekend.]

                                        Be seeing you,

- -- 
Norman.Walsh@Sun.COM    | Nothing will ever be attempted, if all
XML Standards Architect | possible objections must be first
Web Tech. and Standards | overcome.--Dr. Johnson
Sun Microsystems, Inc.  | 
Version: GnuPG v1.2.2 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.7 <http://mailcrypt.sourceforge.net/>

Received on Friday, 1 August 2003 16:29:32 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:00 UTC