[Modeling] Comments on interaction diagrams

Dr. Hong Zhu hong.zhu1@btinternet.com
Tue, 11 Mar 2003 14:27:11 -0000

Dear Jim,

I am reading the document on Interaction Diagrams and trying to make some
comments. I have only read a few pages, but I have a number of questions.
Many of the questions are related to the 'meta-model'. For example, whether
or not agents and objects can mix (i.e. both appear)  in the same sequence
diagram? If so, are there any restrictions on what objects cannot do while
agents can? Is sending messages the only way that agent communicate with
each other? Such questions cannot be answered satisfactorily without a
meta-model of agent-orientation, and they will have a significant impact on
the design and specification of sequence diagrams' syntax and semantics.

>From the emails we had earlier and the information on the AUML web, it seems
to me that we haven't got a meta-model for AUML. May I suggest that we set
up a project on meta-model? As you requested, I am looking into the
meta-model. I have read the papers that you suggested and am now reading the
very thick document of UML 2.0 and re-reading UML's earlier versions. I feel
that setting up a meta-model is not a trivial job at all. It deserve a
separate project and the project should have a high priority.

Another issue that I think is global (i.e. it is not just about the document
on interaction diagrams) is the 'style' of the document. There are two types
of information contained in the docment: background information (including
the history and rationale of the design of the language), and the
specification of the language. For example, the first paragraph on page 2,
is more about the background. I would suggest a different approach. That is,
to separate these two types of information in two documents: one as the
specification of the language, and the other as a complimentary to explain
the background, and to give the rationale of the design of the language,
etc. These two documents can be linked together through hyperlinks embeded
into electronic versions and hosted on the web. Therefore, the specification
document can be short and concise. UML 2.0 proposal is already over 800
pages. We need to think very hard about how to reduce the length of AUML
documents. Otherwise, AUML could be too lengthy to handle and put people

Best regards,