Lists Home |
Date Index |
- From: Jonathan Borden <firstname.lastname@example.org>
- To: Tim Bray <email@example.com>,"Simon St.Laurent" <firstname.lastname@example.org>, email@example.com
- Date: Fri, 29 Dec 2000 22:01:56 -0500
Tim Bray wrote:
> The problems you are (and many others) have pointed out repeatedly
> is that there are large unresolved problems about sharing semantics
> and schemas and other related resources, which are highlighted but
> not solved by XML namespaces.
> I (and I suspect everyone else who helped build the spec) agree
> that these are important and serious problems. I see no inconsistency
> between this position and the position that the spec does what it
> says it does, especially considering the evidence of its widespread
> implementation. -Tim
I beg to differ (actually I don't but I wanted to insert that phrase into
this thread :-) Actually I think that all too often the term "XXX is broken"
or "YYY fails" is used when "XXX" or "YYY" solve only part of a problem, and
perhaps the part that they set out to solve.
Someone recently remarked that nothing in the URI syntax requires or
results in unique identifiers being created. True, but that is not the
point. The point is that it is *possible* to create unique namespace names
using URIs, and certain facilities layered on URIs e.g. incorporation of
network authorities or the presence of the "urn:uuid:" scheme ease certain
aspects of namespace partitioning. The use of URLs as namespace names leaves
open the *possibility* of referencing resources.
So actually I agree. XML namespaces solve the problem it set out to solve
yet because it is built on URIs leave open the opportunity to solve other
problems, involving semantics, ontologies and schemata, that are as yet
unsolved. It is these issues that we need to discuss and thrash out.
The Open Healthcare Group