[Modeling] Interaction Diagram specs - first draft

James Odell email@jamesodell.com
Wed, 05 Mar 2003 08:57:47 -0500


Marc-Philippe,

I agree with Marc-Philippe that we do not want to cause confusion.  At the
same time, I am further wondering if we might find it useful to use *both*
the Sequence Diagram and the Interaction Overview Diagram to represent FIPA
interaction protocol diagrams.  In my experience, both can are need to bring
a clarity and formalism to the IPs that we have lacked in the past.
Furthermore, I have seen some people use Communication Diagrams to express
IPs.  In such a case, using "Interaction Diagram" to name a set of diagrams
is not so incompatible.  Under my proposal, Sequence Diagrams (as a name and
kind of diagram) will still exist, along with  the Interaction Overview
Diagram, Communication Diagram, and the Timing Diagram.  "Interaction
Diagram" is only the name of this "family" of diagrams.  So, one can refer
to the family name or the names of the individuals -- whichever provides
more clarity.  Does that make sense?  (Also, since a few 100,000 users of
UML will be using the same terminology as described above, they will have an
easier time moving to AUML.  Or, at least UML and AUML will be in harmony in
that one respect.)


-Jim




On 3/5/03 8:31 AM, "Marc-Philippe Huget" indited:

> Hi all,
> 
> James Odell wrote:
> 
>> Formally, this was referred to as the Sequence Diagram specification.
>> However, UML 2.0 now uses the term ≥Interaction Diagram≤ to refer to a
>> family of diagrams ? the main one of which is the Sequence Diagram.  The
>> remaining ones are called the Interaction Overview Diagram, Communication
>> Diagram, and the Timing Diagram.  95% of the Interaction Diagram work is on
>> Sequence Diagrams, while the Interaction Overview Diagram is vital to
>> Sequence-Diagram usage.  Therefore, I propose that we change the name of the
>> spec.  (The remaining communication diagram, and the timing diagram portions
>> we can choose to omit or not; they are very short.)  Please let me know
>> whether or not you agree with this change.  If I do not hear from you by 8
>> March, I will assume that either you agree or do not care.  Although, I
>> would like to hear explicitly, if possible.
> 
> Well, it would be difficult for me to say I disagree with this proposal since
> I
> use as much as possible UML 2.0 in this specification. My concern is for
> end-users in a near future. At Barcelona, a guy asked me why we use the term
> protocol diagram to represent interaction protocols in AUML since he knows
> sequence diagrams and not protocol diagrams. If we think FIPA AUML is indebted
> to AUML, we have several names and the new one will need time to be accepted
> since UML 2.0 is so new that a lot of people don't know it exists. Moreover,
> it
> is the problem of consistency for people reading  _old_ papers on AUML and new
> specifications.
> 
> Cheers,
> Marc-Philippe
> 
> --
> Marc-Philippe Huget
> 
> Agent Applications, Research and Technology Group
> Department of Computer Science
> University of Liverpool
> Chadwick Building, Peach Street
> L69 7ZF Liverpool
> United Kingdom
> 
> email: mph@csc.liv.ac.uk
> http://www.csc.liv.ac.uk/~mph
> 
> 
> 
> _______________________________________________
> Modeling mailing list
> Modeling@www.fipa.org
> http://fipa.org/mailman/listinfo/modeling