Re: MetaModel description

[Duane Nickull <duane@xmlglobal.com>]

David:

Some of your questions answered inline:

"Welsh, David" wrote:
> 
> Anders,
> I thought at the last BP/CC joint meeting in Tokyo, on the laast day, there
> a
> discussion and there were basic questions during Karsten's presentation that
> a
> more fundamental question the BP/CC group had revolved around
> "should this specification metamodel layer even exist in the first place ?"

I did a similar guide back in Orlando to begin defining the groups
responsibilities (nicknamed "Duanes' World").  Karstens metamodel is
insightful in some areas but does not fully address all the technical
mehcanisms of ebXML.  It is useful as a reference tool but may not be
used as a normative reference for the TA spec.

> 
> Looking at some of the notes from the Tokyo meetings in the specification
> metamodel *proposal* sent out, a key sentence that stands out goes "Use of
> the
> methodology metamodel is optional but recommended".
> Is that true ?
TA has not endorsed it officially.  It is interesting reading and
Karsten has done a lot of work but it is not a requirement to use.

> I looked in a latest version of the TA and I didn't see mention of
> 'optional', rather the opposite message came across to me.

This version of TA is now depracated.  We found this mistake in several
instances where sections referenced the "ebXML metamodel" yet no
definition was ever included as to what the "ebXML metamodel" was.

> I thought I also clearly heard from Ray Walker to the BP/CC group in Tokyo
> that the Exec/Steering committee has as one of it's decisions yet to take on
> the vote on if UMM is mandated yes or no.
> We have adopted the new UMM in the BP team as what we use to work with.
>>>>>

Personally (not as a TA member), I see no harm in the UMM existing and
it is probably very useful for BPM.  
 
> Is the proposal suggesting the BP methodology layer of ebXML is optional ?

BPM is delivering a starter set of Business Processes along with a well
documented methodology for modelling business processes.  It is very
important becuase we want tobe able to compare other business models and
processes that "newcos" who adopt ebXML may create.  It is very
important that any modelling follow a consistent methodology.  UML has
been named as that methodology.

The BPM layer is not optional becuase BPM must use the modelling
methodology to derive the core set of business processes.

However,  for a company using ebXML methodology, modelling IS optional. 
We cannot dictate that all ebXML participants start their business
transactions by modelling business processes.  Therefore,  modelling is
optional but if modelling is done,  UML must be used in a method that is
consistent with the BPM.


> Also how does the specification metamodel proposal help towards common
> business process
> alignment and further support the fundamental business goals of ebXML ?

The metamodel itself does not fully address this however, the issue of
common business process alignment is an ugly one.  How do we compare two
business process models without human intervention?  It has been
suggested that  by using a consistent methodology to derive XML syntax
from UML sequence (for process flow) and class view diagrams (for
business information), that we can then use an application to compare
but I don;t see this as an implementable solution.

The fundamental business goals of ebXML would suggest to me that
modelling is an important part of the process to ensure that the needs
of business are being met.  We also have a goal to include SME's.  SME's
will not likely have the resources to perform or even comprehend
business modelling and sunsequent XML conversion.

Cheers!

Duane Nickull