RE: [ubl] Discussion of substitution groups

["CRAWFORD, Mark" <MCRAWFORD@lmi.org>]

> And at the risk of saying too much,
> a look at the alternative:
> XSD without derivation looks to me
> like XML without eXtensibility, a
> misnomer. It would put us back to the
> days of csv and fixed width (which some
> still argue is where they'd rather be :-(  ).


The Extensibility in XML comes from being able to create multiple
vocabularies from a single syntax - not from recasting types.  That is a
feature of XSD.  
> 
> XML does provide a major overhead
> to developers and those who fund them.
> I'd argue then that without the promised facilitation of 
> controlled extensibility (without the namespace change seems 
> to me to be without the control) there is little return on 
> investment in terms of software/standards features.

The beauty is not in using a feature of XSD, rather in using XML
end-to-end.
> 
> If XSD is the de facto way to use XML
> then I'd argue that it seems more and
> more that substtution groups are
> becoming the de facto way to use
> XSD to provide eXtensibility in XML,
> especially when faced with a library
> standard such as UBL.

This seems rather faulty logic to me.  Just because a feature is
available, doesn't mean we have to use it. Or perhaps we should use
<any> since that will give us unlimited extensibility.   We are managing
from core components - ABIEs and BBIEs, not from a library of types.  If
you really want to manage by types, then perhaps you should look at a
local solution that is based on some other data modeling methodology.