[ebBP] 11/10/2004: Schema #7 - ProcessSpecification andProcessSpecificationType

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

Schema Work Item #7  ProcessSpecification and ProcessSpecificationType

In Monday's teleconference, we discussed briefly Steve Capell's question 
about the version attribute on ProcessSpecification and 
ProcessSpecificationType. [1]
He asked if the version was related to the BPSS schema definition of 
which the instance complies or the BPSS instance related to the specific 
process being described. [2] We do have an existing v3.0 Work Item 68 on 
artifact versioning. I would suggest we consider that any recommended 
decision support future efforts. Here is a short summary so we can 
decide on actions in call 15 November.

Current syntax:
    <xsd:element name="ProcessSpecification" 
type="ProcessSpecificationType">
        <xsd:unique name="ProcessSpecification-ID">
            <xsd:selector xpath="."/>
            <xsd:field xpath="nameID"/>
        </xsd:unique>
    </xsd:element>

    * If we ask by name for a BPSS instance in a registry, how do we
      know that we have the same one if we don't have a version. 
      Usually a URI includes the version. Complicated with different
      registries. Two URIs are not necessarily unequal if from two
      different registries. - Yunker
    * Does the use of XInclude create any nuances in a version? [4] - Dubray
    * Note that we can't reliably put a version element on metamodel
      elements because a Binary collaboration could remained unchanged
      but a mere change in a BT definition should trigger a change of
      version (I guess I am arguing for a version in Process then...) -
      Dubray
    * Need to consider in line with other previous discussions about
      Name and NameID. [3] -Martin

In Monday's 8 Nov call, we briefly discussed if they could be optional 
but that other residual effects raised by John Yunker (see above).

Let's discuss on the list and come ready to resolve on 15 November. Thanks.

[1] Reference, 29 October 2004: 
http://www.oasis-open.org/archives/ebxml-bp/200410/msg00136.html (original)
[2] We assumed the latter as the nameespace carries the schema version. 
Reference resolution to Work Item 7.
[3] Reference Work Item 7, 
http://www.oasis-open.org/archives/ebxml-bp/200404/msg00125.html
[4] This is kind of odd to put a version on the <process> element for 
instances. With include and all, it is going to be hard to have a 
meaningful version there.