OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

ubl-lcsc message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [ubl-lcsc] UBL-Schema 0p80 draft (alpha edition)


Thanks for the review.  My discussion remarks are found
below.  Certainly, more views and comments are most welcome.


On Mon, 23 Jun 2003, Tim McGrath wrote:

>>Here are my findings when i reviewed these...
>>
>>>A. Application of NDR rules and guidelines
>>>
>>Do these follow the global versus local rule?  I dont know but to me
>>they look like the 0p70 set - which didn't, so we neeed some technical
>>advice here.

If I've heard rightly, the London meeting settled back on 
global elements and global types.  All local stuff was not
admitted.  Which basically means the way it was done in
0p70 was mostly preserved.  0p70 schema set didn't follow
some NDR rules but did follow other NDR rules.

Most notably and probably in relation to this topic of 
discussion, the reusing of types were very much "ignored" in 
0p70 in that very little reuse was made.  Many types of
exact same structure were just re-iterated in the model.
As a consequence, the generated schemas were also re-iterated,
resulting in 92 types defined within reusable.

In this 0p80 alpha, due to a lot more reuse of basic types 
in the spreadsheet model, only 56 types are defined within 
Reusable, a tremendous reduction of new type definitions that 
contributed to the shrinking of Reusable's size (your final 
comment about the reduction of Reusable schema size).

If I recall rightly, reuse of global elements were also 
first found in the 0p80 alpha schemas when in the earlier 
0p70 schemas, no global element was reused due to re-iteration
of new types every time the same structure was needed.





>>>B. Structuring of Schemas
>>>
>>I am not sure why qualified ASBIEs are not referencing their "Type"
>>directly.  for example, why do we need to define a reference for
>>"SendFromAddress", that then just says "type=AddressType"?  I know this
>>was done in 0p70 and maybe it is part of the NDR debate, but does anyone
>>see the advantage in splitting the defintions into two parts of the schema?

To support the NDR global element rule, it becomes necessary not
to define elements directly inside complexTypes, but to refer them
to a globally referenceable element declaration.  This allows
any element to be potentially reused, though as you've probably
observed, not all elements are reused.  But this is just adhering
to NDR global element/type rule.




>>>C. Core Component Types
>>>
>>The Core Component Types schema is not able to find the Type definitions
>>(e.g. AmountType) because the reference wants to have the namespace
>>(e.g. cct:AmountType)    - at least in XML Spy.

Yes, you're right, thanks.  Stephen Green also pointed that out earlier.
I'll add this change to my version for the next draft.




>>DespatchAdvice still has some recursion (not a schema problem but a
>>modeling one).  A DeliveryRequirement has OrderLine, OrderLine has
>>LineItem and LineItem has DeliveryRequirement ad infinitum.

I'll follow this up in the next mail on this topic.





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



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]