Re: [ubl-ndrsc] UBL NDR SC Minutes PM 28 April 2003

[Chin Chee-Kai <cheekai@softml.net>]

We only have the following in 0p70 to work/decide on:

----------------------------------------------
Core Components Types
Reusables
7 document types (Order, Invoice, etc).
----------------------------------------------  (*)

I agree on a judicious mix of both approaches and not get 
stuck in a religious debate of local-is-better-so-everything-is
local, or global-is-great-so-everything-is-global.  Of course,
what's "judicious" remains to be determined.  Everything we
find great now could only be applied at this point in time
with any certainty.  

So assuming a mixed approach, some  stuff will go local and 
the rest global, and we might need an on-going registry of 
some sort to set criteria as to what becomes admitted into 
the global types, and to maintain the evolving introduction/
aging/ retirement of those global types in the registry.
Contextualised UBL schemas or instances could find use of
new types in the registry since that won't affect the
specifications set by UBL as they are contextualised.
But as with any running of registry, who runs it, who 
decides what goes in and out, etc may be another long 
decision making process that I won't get into now.

We can start to be specific about what types and what
elements go local/global by looking through (*).



Best Regards,
Chin Chee-Kai
SoftML
Tel: +65-6820-2979
Fax: +65-6743-7875
Email: cheekai@SoftML.Net
http://SoftML.Net/


On Tue, 29 Apr 2003, Eve L. Maler wrote:

>>I should be clearer: I was assuming that all types would continue to be
>>global, and was proposing that a carefully selected set of elements
>>would be local instead of global.  Does that still sound interesting?  I
>>don't think we've ever seriously considered anonymous types...
>>
>>	Eve