Hi, Le dimanche 19 mars 2006 à 23:25 +0200, Henri Sivonen a écrit : > On Mar 19, 2006, at 19:30, Bjoern Hoehrmann wrote: > > > In Schema 1.1 it is not possible for a xsd:string to be no xsd:anyURI. > > Wow. > > "The ·lexical space· of anyURI is finite-length character sequences." > It says.
The reason for that is that some well known URI schemes (WedDav for one)
don't respect the RFC that defines the syntax of URIs and that it had
been considered that being too strict would create more problems than
extending the lexical space of xsd:anyURI.
> So as far as RELAX NG goes (no infoset augmentation), the whole type
> is useless, right?
Note that although xsd:anyURI has the same lexical space as xsd:string,
its whitespace handling is the same than xsd:token. That makes a
difference when you define enumerations...
As far as validation only is concerned, xsd:anyURI is thus strictly
equivalent to xsd:token (and not to xsd:string) and could be considered
useless.
That being said, RNG schemas are also often used to document
vocabularies and when it's the case, using xsd:anyURI gives a clear
indication that a data is supposed to be used for a URI (or IRI).
Hope this helps.
Eric
--
GPG-PGP: 2A528005
Freelance consulting and training.
http://dyomedea.com/english/
------------------------------------------------------------------------
Eric van der Vlist http://xmlfr.org http://dyomedea.com
(ISO) RELAX NG ISBN:0-596-00421-4 http://oreilly.com/catalog/relax
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------
signature.asc
Description: Ceci est une partie de message numériquement signée
