Re: The range of the HTTP dereference function

 Dan, Tim,
 this is a relative outsider's view:

 Any string can be used to identify anything. Any URI is a 
string. But,

 on the other hand, HTTP has quite well-defined semantics working
with resources represented or updated by documents, and http:  
scheme URLs identify these resources, thus referring to the
document representations.

 Indeed, in XML Namespaces, URIs are used as strings, AFAICS the 
URI form is only so that ensuring uniqueness is much easier. 
Therefore I understand namespace names not as URIs but as strings 
with syntax equal to that of URIs.

 So while a car can be referred to by any string, using an HTTP 
URI for that is IMO mistaken.

 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)

On 25 Mar 2002, Dan Connolly wrote:

 > On Tue, 2002-03-19 at 08:50, Tim Berners-Lee wrote:
 > > Dan,
 > > 
 > > Here is my argument the HTTP URIs (without "#") should be understood as
 > > referring to documents, not cars.
 > [...]
 > > Why do you want to extend the range of http URI dereference to cars?
 > >From what I can tell, folks agree with me that http URIs
 > already can refer to cars.
 > After all: Web Architecture is minimally constraining;
 > the burden of persuasion seems to be with you, TimBL,
 > since my picture of the Web is less constraining.
 > Does anybody agree with TimBL?
 > If so, please speak up.
 > Otherwise, I'll devote my attention elsewhere for a time.

Received on Wednesday, 27 March 2002 07:18:09 UTC