RE: Suggestion for ebXML Knowledge Dissemination

[Christian Huemer <ch@ifs.univie.ac.at>]

> I think Mr. Kammerer is just what the metamodel document
> needs right now: a critical reader.  His main criticism
> is that the UML diagrams are unreadable, especially
> the first big one:
> 
> >Yes, I realize use of UML was "mandated.  But pretty much all the UML
> >diagrams give me a headache - perhaps I need to take the UML for Dummies
> >class.  And the ebXML Business Process Metamodel on Page 4 of the BP
> >First Draft is out of this world;  I would call it spaghetti-code. 
> 

i do not think that the main design goal for the bp meta model, is
that everybody participating in ebxml has to understand it. the
meta model must be able to capture all business semantics within
a business transaction. 

not every word user has to understand the word file format
before typing a letter. you can use your ordering application without
understanding the data structures. modelers can use a uml modeling tool
without being aware of the uml meta model. what these users allowed
to do is constrained by the tool/application. when you want to teach 
somebody in UML, the probably worst thing you can do is to show 
and explain him/her the meta model behind UML. how many UML books have
you seen presenting the UML meta model (although the content is conform
to the meta model).

in my opinion we need tools allowing modelers to define ebxml conform
business models (which means that the models are conform to the bp
meta model). but before we can attract software tool providers to
build such a tool, we need a rather stable meta model. similarly, a
new version of the UML meta model is standardized, before tool vendors
incorporate it into their tools.

good tools will also guide (by the application itself) the modeler in
producing a ebxml conform model. this means that they incorporate
a methodology in producing an ebxml model. so what we need is a first
proposal on a methodology which shows how - by following the methodology -
an ebxml compliant model is produced. this will help the "dummies" to
understand the function of the meta model. than we have to verify if
existing methodologies (edifecs/rosetta net, swift, oag, tmwg, ...) can
deliver a ebxml compliant model. Even more important we must evaluate 
whether the output (ebxml model) is the same regardless of the 
methodology used or not. i strongly believe not. if i am right, i think
we need a common methodology.

nevertheless, i like bob's approach of defining packages according
to a narrative flow. i think this will also help in giving a first
guideline on how to produce an ebxml compliant model.

christian


=======================================================================
= This is ebxml-bp, the general mailing list for the ebXML            =
= Business Process project team. The owner of this list is            =
= owner-ebxml-bp@oasis-open.org                                       =
=                                                                     =
= To unsubscribe, send mail to majordomo@lists.oasis-open.org with    =
= the following in the body of the message:                           =
=      unsubscribe ebxml-bp                                           =
= If you are subscribed using a different email address, put the      =
= address you subscribed with at the end of the line; e.g.            =
=      unsubscribe ebxml-bp myname@company.com                        =
=======================================================================