RE: httpRange-14 , what's the problem

From: Jonathan Borden [mailto:jonathan@openhealth.org]

>1) If HTTP URIs necessarily identify documents, what does

>http://www.w3.org/2001/XMLSchema

>identify?

According to the browser:

"XML Schema
13 February 2001 ... "

Here is the rub:

"The basic fallacy is that you can make the system general by introducing a second level - a new set of attributes, properties, or whatever, which allow you to refer to the metadata of something separately from the thing itself. These systems either turn out to be just limited 2-level systems (like XML and DTDs) or have to be extended to be recursive in some way later on such that in fact the two levels become unnecessary."

It isn't a fallacy so much as a category error.

If you generalize a system, it is no longer the same system.  A two level system happens when a 
system is identified/categorized in terms of another system.   That really isn't a problem until 
one loses track of which system one is working in at the time.  The browser has one assignment 
for the thing identified, and others are possible, but only in other systems.   I don't think 
it useful to try to make identification (uniquely naming and maintaining a unique name) and 
categorization (labeling a thing by picking a name from a preexisting set of labels) work from 
the same label.

Berners-Lee does have the practical answer:  make it mean a unique document when and 
if dereferenced.   I don't believe that confuses people.

len

Received on Monday, 29 July 2002 14:07:09 UTC