class vs. metaclass (was: Re: associations defined in a metamodel



Date view Thread view Subject view Author view Attachment view

From: Dirk Riehle (dirk@riehle.org)
Date: Fri 17 May 2002 - 11:40:15 BST


A short question please, on how the class vs. metaclass issue is considered:

Do you consider the concept of Association as defined in UML a class or a 
meta-class?

In strict metamodeling, you would probably consider it a meta-class, as you 
reserve application models to contain classes only.

In non-strict metamodeling, and I believe this where UML 2.0 is headed, I 
think Association will be a class, literally an instance of the Class concept.

Thanks,
Dirk



At 5/15/2002 +0100, Andy Evans wrote:
>Simona,
>
>It depends what you mean by instantiated!
>
>Do you mean:
>
>1) An instance of the meta class association, association end, etc.
>
>OR
>
>2) An instance in the sense of links, link ends, etc. (i.e. the semantic 
>domain
>of an association)
>
>If the latter, then there is no similar concept to multiplicity in the 
>semantic
>domain. Rather, there is a check made that the number of links between
>objects conforms to the multiplicity of the association ends.
>
>Andy
>
>----- Original Message -----
>From: "Simona Bernardi" <bernardi@di.unito.it>
>To: <puml-list@cs.york.ac.uk>; <bernardi@di.unito.it>
>Sent: Wednesday, May 15, 2002 4:45 PM
>Subject: associations defined in a metamodel
>
>
> > Hello,
> >
> > I have a question about how associations defined in a UML meta-model are
> > instantiated.
> > What happens to the multiplicities of an association  defined between
> > two meta-classes ?
> > For example an association  "ass" is defined between two metaclasses A
> > and B,
> > with multiplicity * (not explicitly defined):
> >
> >         ass
> > |A| -----> |B|
> >           *
> >
> > An instantiated model can be made of  a class A' (instance of the
> > metaclass A)
> > and a class B' (instance of B) and relation ass' between A' and B'.
> > Does the multiplicity of ass' have to be a specific number or it can
> > still remain unspecified with "*" ?
> >
> > Thanks,
> > Simona.
> >
> >
> >
> >
> > To remove yourself from this list please mail 
> puml-list-request@cs.york.ac.uk
> > with a message containing the word "unsubscribe".
> >
> >
>
>
>
>
>To remove yourself from this list please mail puml-list-request@cs.york.ac.uk
>with a message containing the word "unsubscribe".

----
For work etc. see: www.riehle.org

Date view Thread view Subject view Author view Attachment view