Lists Home |
Date Index |
Bart Schuller wrote:
>On Mon, Apr 11, 2005 at 06:37:12PM +0200, Jan Algermissen wrote:
>>and I have never ever had any problems with that decision. I believe
>>that a technology itself gets to define what their 'elements' are and
>>not whoever uses the technology.
>>A nice example of this is XTM (Topic Maps applied to XML/Web) which
>>implicitly makes the assumption that URIs allways identify Documents
>>(aka 'addressable subjects') and NEVER!! abstract concepts (aka
>>'non-addressable subjects'). How can a technology (Topic Maps) that
>>*uses* terms and infrastructure of another technology re-define the
>>terms? Makes no sense to me.
>It makes no sense because it isn't true. It explicitly *DOES* use URIs
>in precisely these two ways.
REST's definition of resource identifier (URI) does not support the use
of URIs in two different ways. It is an illusion that
you can change the semantics of HTTP by providing an alternative
definition of URI.
Would you ever argue about using object identifiers in any piece of OO
application in two different ways?
>See for example
>Now from what I've read about RDF it seems that too uses URIs in these
>different ways, but fails to make the distinction, so you can never be
>sure whether a statement talks about a document or about the subject of
Consultant & Programmer