Re: New draft TAG Finding on The Self-Describing Web

> I think the core of the concept "information resource" is that it's
> a resource that we are willing to identify with its representations.
> A facsimile of the First Folio, or even a faithful textual transcription,
> just *is* the First Folio, as much so as any of the 40-odd surviving
> printed copies; but nobody would think that a picture of Shakespeare
> really *is* Shakespeare, although it's certainly a representation of him.
>   
Is a picture of Shakespeare a "representation" or a "description" of 
him? IMHO, the essential difference between an information resource and 
a non-information resource is only the former can have a 
"representation" in the web.

What I found a bit of troubling about the draft is this paragraph

"Note that the above wording takes account of the fact that HTTP 
headers, such as Content-Type, are considered to be part of the resource 
representation, even though they are not part of the HTML entity-body 
content; indeed, ...."

This seems to tie a resource tightly to a network protocol.  Should a 
HTTP resource be "dereferenced" via another protocol, say FTP, their 
representation would be different by the above wording. Headers are 
metadata about the representation and should not be considered to be 
part of the representation.

Xiaoshu

Received on Wednesday, 30 May 2007 18:50:49 UTC