QNames in content

Hi,

as noticed recently on www-tag[0], QNames in content aren't consistently used 
accross specs. Some (like XML Schema, 3.15.3.2.2.1) have them use the default 
namespace, others (like XSLT, 2.4) don't.
Question one:

Does the following snippet adequately capture a default-namespace-insensitive 
QName (the regex might be slightly off as I'm not fully familiar with XML Schema 
regexen, but it's "NCName:NCName" to ensure it can only be a QName if it has a 
prefix):
<simpleType>
  <union>
    <simpleType>
      <restriction base='QName'>
        <pattern value='[_\i]([\c]-[:])*:[_\i]([\c]-[:])*'/>
      </restriction>
    </simpleType>
    <simpleType>
      <restriction base='NMTOKEN'/>
    </simpleType>
  </union>
</simpleType>


Question two:

It's too late to decide on only one option and make all specs consistent on this 
point, but wouldn't it be a good idea for a WG (XML Core?) to publish short and 
simple guidelines for when QNames in content are appropriate and when which 
option (defaulting ns or not) is best?
[0]http://lists.w3.org/Archives/Public/www-tag/2003Jan/0096.html

--
Robin Berjon <robin.berjon@xxxxxxxxx>
Research Engineer, Expway        http://expway.fr/
7FC0 6F5F D864 EFB8 08CE  8E74 58E6 D5DB 4889 2488