[Fwd: Re: Moberg 9/20/2005: UBL Instance and Spec Reference]

[Monica J Martin <Monica.Martin@Sun.COM>]

This is the inwork update to the UBL-ebBP XML instance. Stephen Green, 
Dale Moberg, Dean Hemopo and I are collaborating to determine 
upgrades/updates to the instance including use of external roles and 
effective use of nameID. Stephen is also considering some other modular 
process fragments for the SBS (Small Business Subset). Any comments 
welcome as we collaborate with Dean on this instance. Thanks.

--- Begin Message ---

• From: Stephen Green <stephen_green@bristol-city.gov.uk>
• To: dmoberg@cyclonecommerce.com, stephen_green@seventhproject.co.uk,Monica.Martin@Sun.COM
• Date: Wed, 21 Sep 2005 15:55:18 +0100

Dale

Greetings. Please find attached a zip file (renamed with
.zzz extension) with corrections (.xml) and change log (.rtf
with changes made in bold/italics/underlined)
for the UBL example BPSS instance (note I also changed
the schema name to 2.0.1).

All the best

Steve


>>> "Dale Moberg" <dmoberg@cyclonecommerce.com> 21/09/05 15:17:11 >>>
Thanks for your clarification.

I am getting closer to understanding the problem now.

We have called these non-schema expressible constraints
"well-formedness" rules because they capture semantic relations that
cannot be expressed as schema constraints. In your case it is something
like "type" information on IDs and IDREFs. You are quite correct that
syntactically any ID value can serve as a value for an expression of
type IDREF. The Business Logic of the example was not separately
documented and I am afraid I did not closely review it. Also, we did
have to update the Performs elements in the example because we changed
attribute names at the last minute. So I may well have introduced an
editorial error when working on those changes.

I will look over the example from the perspective you have provided. Did
you previously send the corrections you propose? A quick search failed
to find them but I may very well have not looked in the right places.

Dale Moberg

-----Original Message-----
From: Stephen Green [mailto:stephen_green@seventhproject.co.uk] 
Sent: Wednesday, September 21, 2005 5:38 AM
To: Monica J Martin; Dale Moberg
Cc: Stephen Green
Subject: Re: Moberg 9/20/2005: UBL Instance and Spec Reference

Monica, Dale

I found the problem - it is that
nameID="Shipper" is used in

<ExternalRoles>

<BusinessPartnerRole nameID="MeSeller" name=""/>

<BusinessPartnerRole nameID="MeBuyer" name=""/>

<BusinessPartnerRole nameID="Buyer" name=""/>

<BusinessPartnerRole nameID="Seller" name=""/>

<BusinessPartnerRole nameID="Shipper" name=""/>

<BusinessPartnerRole nameID="Receiver" name=""/>

</ExternalRoles>


and so the use of "Shipper" for an ID is valid anywhere in the instance.
However
that is because the schema allows * any *
ID to be used anywhere where an ID is
required (perhaps a limitation of XSD).
So the XML is valid but the business logic
of the instance isn't and that is because such
business logic cannot be (or at least, is not)
enforced by the schema.

This means that just the instance should be
corrected of course.

All the best

Steve


----- Original Message ----- 
From: "Monica J Martin" <Monica.Martin@Sun.COM>
To: "Dale Moberg" <dmoberg@cyclonecommerce.com>
Cc: "Stephen Green" <BRITSDG@bristol-city.gov.uk>; "Stephen Green"
<stephen_green@seventhproject.co.uk>
Sent: Tuesday, September 20, 2005 7:00 PM
Subject: Re: Moberg 9/20/2005: UBL Instance and Spec Reference


> Dale Moberg wrote:
>
> >I tried the example as sent against the schema as sent (or referred
to)
> >and the example validated in oxygen (5.1), need to update.
> >
> >I also tried in xmlspy 2004 rel 3, and the instance validated.
> >
> >
> mm1: Then I would suggest we manually check nameID and its use in
> performs, and understand:
>
>     If we need to apply any specific constraint that isn't currently.
>     If we missed anything in the definition of Performs (and Performs
>     Type). Currently the nameID is xsd:ID on nameID of name attribute
>     group. Performs doesn't use optname attribute group that allows
for
>     optionality.
>
> What do you both think? Thanks.
>
> >Dale
> >
> >-----Original Message-----
> >From: Monica J Martin [mailto:Monica.Martin@Sun.COM] 
> >Sent: Tuesday, September 20, 2005 10:10 AM
> >To: Dale Moberg; Stephen Green; Stephen Green
> >Subject: Moberg 9/20/2005: UBL Instance and Spec Reference
> >
> >Here are the actual Committee Draft v2.0.1 specification (in Word)
and
> >the UBL instance used (.xml) attached.
> >
> >Spec reference (in .zip):
>
>http://www.oasis-open.org/apps/org/workgroup/ebxml-bp/document.php?docu 
m
> >ent_id=13905
> >Schema reference (in .zip):
>
>http://www.oasis-open.org/apps/org/workgroup/ebxml-bp/document.php?docu 
m
> >ent_id=13907
> >
> >Stephen and Dale, could you please advise where the nameID is used
> >inappropriately so we can ensure the instance validates and logically
> >so.
> >Thanks.
> >
> >
>
>

PKZIP (compressed) files

--- End Message ---