[Modeling] RE: Comments on interaction diagram modeling
Dr. Hong Zhu
Thu, 20 Mar 2003 17:02:29 -0000
Sorry for joining the discussion so late. Reading the emails passed around,
I feel that there are a lot of issues involved in the discussion. However,
the most important thing is what are the 'requirements' of such an
interaction diagram and whether the current version meets the requirements.
Let's have a more concrete example and see how interaction diagram to
represent a protocol. Consider the procedure that UN security council passes
a resolution. It seems the protocol consists of the following steps:
(1) At least one member of the security council submit a proposal;
(2) By a given date that the council members all agreed, a vote from the
members must be made;
(3) Each member of the security council can vote either FOR or AGAINST or
(4) The proposal becomes a resolution of UN security council's resolution,
if the majority of the members voted FOR, and no permanent member voted
Here, we have two castes (i.e. agent types): One for security council
members and one for permanent members. Agents of these two castes are
autonomous, but they together form another agent - UN security council.
Can we represent this protocol in the interaction diagram?
----- Original Message -----
From: "Wagner, G.R." <G.R.Wagner@tm.tue.nl>
To: "'Renato Levy'" <firstname.lastname@example.org>; <email@example.com>
Sent: Thursday, March 20, 2003 9:44 AM
Subject: RE: [Fwd: [Modeling] Comments on interaction diagram modeling doc]
> >> protocols are both a design specification tool, but also a
> >> standardisation mechanism. We may need to address these 2 uses
> >> separately?
> > This is precisely the point for the definition of level of
> > abstraction of the design document.
> It's not really about the "level of abstraction" but rather
> about the purpose of the model: standard interaction protocols
> correspond to domain models which you may transform into
> design models (by enriching/refining them). The OMG also
> calls these two different types of models "Computation-
> Independent Model (CIM)" and "Platform-Independent Model
> (PIM)" in their Model-Driven Architecture approach.
> Modeling mailing list