W3C home > Mailing lists > Public > www-tag@w3.org > October 2004

Re: [Fwd: RE: "information resource"]

From: Roy T. Fielding <fielding@gbiv.com>
Date: Sat, 16 Oct 2004 13:27:42 -0700
Message-Id: <D4A64BCC-1FB1-11D9-B3CA-000393753936@gbiv.com>
Cc: <www-tag@w3.org>, Mark Baker <distobj@acm.org>
To: Harry Halpin <hhalpin@ibiblio.org>

On Oct 16, 2004, at 11:06 AM, Harry Halpin wrote:
> I think Mark's point about the medical record can be turned around on 
> its
> head pretty easily. For example, a doctor may write on the note at time
> t1, and this does not ofcourse effect an XML document made of the 
> record
> at an earlier point (t0). However, this can happen in reverse: the XML
> document version of the medical record (t0) could be sent over the Net 
> to
> another doctor, who then writes another note on the XML document (at 
> time
> t2) that *is not* on the original note (t0). Does this mean that this 
> new,
>  updated XML document (t2) is still a "representation" of some
> medical record (t0).

A representation is not some second-class object that doesn't
exist outside of its original relation to the resource.  If the
representation is modified then it is simply a modified representation,
which is a *different* resource in its own right. The modification
has no impact on the original resource unless an action is taken
(directly or indirectly) on the original resource, after which its
later representations will reflect the new state.

For example, in all but the most high-tech hospitals, notes will be
recorded on paper during the day and a clerk or nurse will enter
the accumulated changes into the patient's record at some point
during the day (or night).  Does that mean that the original resource
is not the true medical record of the patient?  Yes, of course -- it
is only the medical record as known by the provider of that resource.
Is that a problem? Yes, and it is well-known in the field of medicine,
which is why the US just authorized implantation of chips for storing
medical records (of a sort).  It shouldn't be too hard to consider
what that would mean if the chip is an RF-capable HTTP server that
allows access to its resources under access control, perhaps
even including resources that provide real-time monitoring of
heart-rate, blood pressure, and blood sugar.  If you want to get
really fancy, consider an embedded insulin monitoring system that
is 99% self-sufficient aside from a configuration/monitoring
interface with a virtual on/off switch.

Those are all "information resources", but there is no effective
difference between that class of resource and anything else that
might be identified by a URI.  Arguing about their existence is
a complete and total waste of time, since any system that relies
on such a distinction is inherently broken and no better off than
it would have been by accepting that resources are resources and
assertions about nouns are usually ambiguous, requiring additional
assertions to differentiate between terms.

What makes a URI unambiguous is the degree of trust you can place
on the naming authority to describe its meaning accurately and in
a way that everyone understands and agrees to, and to maintain that
meaning over all time.  Whether or not information is supplied by
some system in response to a GET-equivalent request for that URI
does not change reliance on the naming authority.

> Even if the original (t0) is modified later (at t1)?
> Is not the XML document (t2) *a medical record* in its own right, 
> albeit a
> different one from both the original(t0) and another modified 
> version(t1),
> in its own right? So I'm not sure how the use of the word 
> "representation"
> has any real meaning in this case, except in the technical sense that
> for the XML copies (assuming they have URIs) an HTTP GET gets you
> something. As the more philosophical use of the "representation" 
> meaning
> "a representation of", I think that use of the word only confuses 
> things.

No, you are simply assuming that the resource is one thing and
then complaining when its representations do not actually correspond
to that thing.  The assumption was wrong.  The data is a representation
of something, even if that something isn't known to the requester.
The reason for the word is because people constantly assume that
performing GET on a URI transfers the resource itself, which simply
isn't the case (even for a simple scheme like "file").

....Roy
Received on Saturday, 16 October 2004 20:28:14 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:43 UTC