Re: [ubl-ndrsc] Bill&Eve's proposal on local/global elements

[Dave Carlson <dcarlson@ontogenics.com>]

----- Original Message -----
From: "Eve L. Maler" <eve.maler@sun.com>

> The following was whipped up by Bill and Eve after we recessed for the
> day.  We hope it's coherent, and will lead to productive discussion
tomorrow.
>

Hi gang,
I wish that I could be there for this discussion tomorrow, but here are my 2
euros.

Using Bill & Eve's example, I constructed a very simple UML class diagram in
Rational Rose.  See the attached GIF image.  It's important to note that this
UML model contains *NO* information specific to XML; it's just an abstract
information model, independent of any implementation syntax.

However, when I import this model into my schema generation tool, I apply a
set of default transformation rules and get an XSD that looks identical to
Eve's example.  See Party-1.xsd attached.  Here is an excerpt:

<xs:complexType name="PartyType">
   <xs:sequence>
      <xs:element name="name" type="xs:string"/>
      <xs:element name="Identifier" type="IdentifierType"/>
      <xs:element name="Address" type="AddressType"
                         minOccurs="0" maxOccurs="unbounded"/>
   </xs:sequence>
</xs:complexType>

My schema generator also includes a configuration switch to generate qualified
names, which was originally intended for generating DTDs that required
globally unique element names.  See Party-2.xsd attached.  An excerpt:

<xs:complexType name="PartyType">
   <xs:sequence>
      <xs:element name="Party.name" type="xs:string"/>
      <xs:element name="Party.Identifier" type="IdentifierType"/>
      <xs:element name="Party.Address" type="AddressType"
                         minOccurs="0" maxOccurs="unbounded"/>
   </xs:sequence>
</xs:complexType>

Although I would NOT recommend generating a schema like this one, it
illustrates Eve & Bill's point that you can derive unambiguous dictionary
entry names by qualifying local names with their scope name.  These qualified
names (e.g. Party.Identifier and Address.Identifier) are identical to those in
Eve's email message.

Because both of these schemas where auto-generated from a generic UML
information model, it would be a trivial exercise to also generate a set of
dictionary entries for a BIE catalog using the qualified names.  If you assume
that the future Core Components and/or BIE models are abstract models similar
to my simple UML diagram, then this one model can generate both XSD schemas
using the simple locally scoped names, plus a set of qualified names used for
other purposes.

Virtually yours,
  Dave Carlson

Party.gif

Party-1.xsd

Party-2.xsd