[relax-ng] Type-assignement requirements from Jonathan Robie

["MURATA Makoto (FAMILY Given)" <EB2M-MRT@asahi-net.or.jp>]

I exchanged some e-mail with Jonathan Robie (a member of the W3C XML Query WG) 
about type-assignement.  I find his e-mail very useful.

> Sure, I'll let you know directly, and you can feel free to forward my 
> emails or quote me.
> 
> The main requirement is to provide type annotations for every element or 
> attribute, perhaps in a separate process from validation. In XQuery, every 
> element or an element has a type annotation, which is a QName. Only 
> elements and attributes have type annotations. Sequences, model groups, 
> etc. do not. If an element does not have a named type, its type annotation 
> is xs:anyType; if an attribute does not have a named type, its type 
> annotation is xs:anySimpleType.
> 
> W3C XML Schema has both named and anonymous types. There is a proposal on 
> the table to create an implementation-defined internal name for every 
> anonymous type when importing a schema, and using that internal name as the 
> type annotation for instances in the data model. I think this is the right 
> thing to do. If you want to do this in a standard way, you can number 
> anonymous types as you encounter them relative to the nearest named type or 
> element, or whatever.
> 
> Makoto wrote:
> >For example, do you want to know which subsequence of elements matches
> >which subpattern?
> 
> No, XQuery would not be able to use that information, and it would be lost 
> in the mapping to the Data Model.
-- 
MURATA Makoto (FAMILY Given) <EB2M-MRT@asahi-net.or.jp>