>> That said, as I've argued before, substitution groups are not
>> as extensible a mechanism as use of xsd:any.
>Sure. Lets have two types in UBL. Any and Thing with unlimited content
>models. Very extensible - and totally useless for setting STANDARDS.
This is of course far different from having x very specific types with clear
meanings that must be present and must be handled, y specific types with
clear meanings that might be present and must be handled if present, and 1
type that contains unknown content that must be ignored if not understood
and can be handled if understood, and that has some strictly defined
attributes to inform any application as to the domain and origin of the not
understood content.
Regards
Bryan Rasmussen
