[Uml-devel] bug

[Jens Krüger]

Am Mittwoch, 27. November 2002 23:44 schrieb Jonathan Riddell:
> > > I have never seen such a thing in UML books. Do you really think this
> > > is necessary? I mean if it is possible to have multiple transition, why
> > > is there fork/join/branch/merge?
> >
> > well... the forj/join is a different beast: those are to represent
> > different threads of execution --> to go to activities which are supposed
> > to happen in "parallel" (fork) or to synchronize the execution at the end
> > of two activities (join)
> >
> > s for branch/merge.. yes.. I suppose you are right:
> > All you can achieve with multiple transitions can be represented with a
> > branch (actually, I think this is the clearer way to do it).. I just
> > think that having the transitions directly from the activities could save
> > some work.
> >
> > UML says that activities can have more than one outgoing transitions as
> > long as they are differentiated by conditions (if [condition a]
> > transition 1, if [condition b] transition 2, else transition 3 )
> > which is precisily what branches do ... again, it would be only for
> > convencience... but it's up to you if you want to accept this.
>

I think it was no good idea to allow multiple outgoing associations from an 
activity. I've never seen this in any literature, but what I've seen is that 
an object is used inside an activity diagram which transports information 
from one to an other activity. But we haven't implemented the occurence of an 
object in an activity diagram, so we should not allow multiple outgoing an 
ingoing associations from and to activities. I think, the idea of the 
singularity of associations is well founded. The designer should use the 
branche/merge and fork/join elements to model. Otherwise you have no 
mechanism to check in what case you use the one and in what case you use the 
other association.

Please go back to the previous behaviour.

Regards

Jens

> I've changed assocrules.cpp to allow for multiple outgoing associations
> from activites.  It would seem strange for state diagrams to allow them
> but not activity diagrams.
>
> Jonathan Riddell
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Get the new Palm Tungsten T
> handheld. Power & Color in a compact size!
> http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en
> _______________________________________________
> Uml-devel mailing list
> umbrello-devel at kde.org
> https://mail.kde.org/mailman/listinfo/umbrello-devel