W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2001

Re: Namespaces wihtout "#" Was: Few CWM Bugs

From: Roy T. Fielding <fielding@ebuilt.com>
Date: Sun, 25 Nov 2001 15:50:24 -0800
To: Piotr Kaminski <piotr@ideanest.com>
Cc: www-rdf-interest@w3.org
Message-ID: <20011125155024.A1587@waka.ebuilt.net>
> That's interesting -- do you claim that there are no resources that can be
> transmitted directly over a network?  What about a resource like "the
> sequence of bits 01101011"?  If you put those bits in a packet, are you
> transmitting the resource or a representation?  If I recreate these bits
> on my client, do I now have your resource, or just another resource that
> happens to have an equal value?  In other words, what is the identity of a
> sequence of bits?

The identity that you, as a naming authority, placed on it.  For example,
they could be "my favorite bits", "the bits I saw last week", or "the
bits that exist at a certain offset within some picture of me represented
as a GIF."  Or simply "the sequence of bits 01101011".  In all cases you
are transferring a representation, because there exists no mechanism in
the universe to transfer the semantics of identification that you expressed,
even for something so simple as "the sequence of bits 01101011".

What you can do is transfer a representation, r, and then transfer a
statement in temporal logic that says, for all times t and t',
  if r  is a representation of the resource R at time t  and
     r' is a representation of the resource R at time t', then

     r == r'

and you can wrap that logic into a resource and refer to it via some
other well-known URI.

That is what your brain does when someone says that the resource R is
just a static sequence of bits that never changes.  That notion of a
resource is separate from the issue of whether those bits are represented
as a sequence of high and low voltages, a sequence of binary digits in
network-byte order, or a sequence of "0" and "1" characters in ASCII.

> These questions may seem pedantic, and I'll grant that we very rarely care
> to talk about such primitive resources directly, but I think they get at
> the heart of the distinction between a resource and its representation.

Yes.  The important thing is that, once the distinction is clear and the
mechanisms of the Web are free to manipulate everything as representations,
the actual semantics of what it means to be a resource is completely
hidden behind the abstract interface, and therefore not limited by any
preexisting notion of what is within the scope of identification.

....Roy
Received on Sunday, 25 November 2001 18:54:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:52 GMT