[Modeling] Modeling an Agent Class- problem with operations:

[Wagner, G.R.]

> I am not understand how UML's way of defining semantics and 
> syntax. Could
> you please tell me how to turn this feature off when the semantics are
> written in the definition of 'class'?

I think you can define a class stereotype "AgentType" where you
suppress the operations compartment with a suitable OCL invariant
attached to this stereotype.
 
> Second, restrict the caller to the object itself does not 
> solve all the
> problems. Method called is the only means of communication 
> between objects.
> If no call to other objects' methods, how to communication 
> with each other?

Jim has pointed out other ways of communication between
active objects (by means of "signals", which are simply
message types).

-Gerd


> ----- Original Message -----
> From: Wagner, G.R. <G.R.Wagner@tm.tue.nl>
> To: Dr. Hong Zhu <hzhu@brookes.ac.uk>; ModelingTC <modeling@fipa.org>
> Sent: Monday, June 23, 2003 06:56
> Subject: RE: [Modeling] Modeling an Agent Class- problem with 
> operations:
> 
> 
> > There seem to be two possible solutions for your problem 
> with operations:
> >
> > 1) turn off this feature (of classes having operations that can be
> > called) in a AUML Profile for domain/analysis modeling
> >
> > 2) restrict the allowed callers of the operation to self (by means
> > of suitable constraints defined for the AUML Profile).
> >
> > Does this help?
> >
> > -Gerd
> >
> >
> > -----Original Message-----
> > From: Dr. Hong Zhu
> > To: ModelingTC
> > Sent: 23.06.2003 10:21
> > Subject: Re: [Modeling] Modeling an Agent Class- register 
> your opinion
> >
> > Let's assume that it can be done this way, then, we must 
> consider agents
> > are
> > special case of objects, because the semantics of class is 
> defined as
> > follows. (the following is taken from Jim's email about 
> UML's definition
> > of
> > class)
> > "-Semantics
> > The purpose of a class is to specify a classification of 
> objects and to
> > specify the features that characterize the structure and behavior of
> > those
> > objects."
> >
> > The question is, then, "Can agents be objects?" I doubt about it,
> > because
> > the semantics of object has this features defined as a part of UML
> > class'
> > semantics: (also taken from Jim's email)
> > "Operations of a class can be invoked on an object, given a 
> particular
> > set
> > of substitutions for the parameters of the operation." Does 
> this give
> > object
> > the freedon to refuse invocation of its operation? I think, a
> > characteristics of agents is that "agent can say no to operation
> > invocation". This is also from Jim's work.
> >
> > Hong
> >
> > ----- Original Message -----
> > From: "James Odell" <email@jamesodell.com>
> > To: "ModelingTC" <modeling@fipa.org>
> > Sent: Saturday, June 21, 2003 7:01 PM
> > Subject: Re: [Modeling] Modeling an Agent Class- register 
> your opinion
> >
> >
> > > On 6/20/03 7:45 AM, "Dr. Hong Zhu" indited:
> > >
> > > >> When you simply use the word Agent as the class name 
> of the class
> > Agent
> > > >> you can produce a meta model in UML. Using UML as the modelling
> > > >> language does not mean that the behaviour of agents 
> necessarily has
> > been
> > > >> limited to OO.
> > > >>
> > > >
> > > > I don't think the semantic definition of UML can be 
> interpreted that
> > way.
> > >
> > > And, I am not so sure that it cannot be.  I guess it's 
> time for us all
> > to
> > > demonstrate it one way or another.
> > >
> > > -Jim
> > >
> > > _______________________________________________
> > > Modeling mailing list
> > > Modeling@www.fipa.org
> > > http://fipa.org/mailman/listinfo/modeling
> > >
> >
> > _______________________________________________
> > Modeling mailing list
> > Modeling@www.fipa.org
> > http://fipa.org/mailman/listinfo/modeling
> >
> > _______________________________________________
> > Modeling mailing list
> > Modeling@www.fipa.org
> > http://fipa.org/mailman/listinfo/modeling
> 
> _______________________________________________
> Modeling mailing list
> Modeling@www.fipa.org
> http://fipa.org/mailman/listinfo/modeling
>