Lists Home |
Date Index |
Tim Bray wrote:
> John Cowan wrote:
> > I do not understand the URIs Good, QNames
> > Bad point of view, since they are plainly isomorphic. Indeed, RDF
> > a two-way mapping between QNames and (some) URIs.
> I don't think so. RDF defines such a mapping, but (for example)
> XSchema, which allows identification of type by qname, does not. I
> might be a little less nervous about qname proliferation if they were in
> fact isomorphic to URIs.
I hear the "QName in attribute values are bad" argument but cannot bring
myself to accept it. QNames are, perhaps, better syntactic identifiers than
URIs -- at least my human eyes greatly prefer them to longhand URIs.
XPath/XSLT demonstrates the success of QNames in attributes. Aside from the
issues surrounding the current XML Schema inability to provide a URI for a
QName type, _reading_ <foo xsi:type="xsd:integer">34</foo> is a heck of alot
better than reading <foo
> > It would be more consistent
> > for you to attack URIs, QNames, and IP addresses, all of which are
> > universal agreements creating global names,
> Well, such universal agreements are expensive. We have two: the IP
> address space and the DNS. We hace a universal naming scheme, the URI,
> that builds on these. Can we please stop and not invent any more?
Fair enough -- which is exactly *why* we need a standard mapping of QNames
to URIs and back as I've suggested in
Now since your are the issue owner :-) if you can solve this problem I'd be
very very happy to show you how RPV can be made -even more readable-
Just allowing QNames rather than URIrefs in the current RDF/XML would fix
alot of the readability issues e.g. as I've suggested long ago:
However I agree that demonstrating the triples explicitly as in RPV or
N-Triples/N3 makes RDF alot more readable for humans - in particular N3's
combination of explicit triples and QNames.