I love it when "simple" questions turn into long, contradictory
discussions.

The issue is:  Can a use case "point to" an actor in a use case
diagram?  Further discussion shows that the intent of the question is
more like:  Can information flow to an actor from a use case?

The best short answer is:  The use case diagram doesn't say that kind of
thing.

The sequence diagram *does* say whether objects in the use case
realization pass messages to an actor.  In this context, the situation
is clearer.  Passing a message to an object implies that the object has
an operation that is a part of the sequence.  Passing a message to an
actor also implies that the actor has an operation that is part of the
sequence.  You can say this about a human actor if you want to, but how
will anybody realize it?  If you are doing system modelling, I advise
against passing messages to human actors.  It would be acceptable in
business modelling, but that wasn't the intent of the original question.

Going back to the use case diagram:  The association between an actor
and a use case does imply <<communicates>> (if I remember correctly),
but it does not indicate a direction.  Of course, you can nudge the
association into saying something more, by using your own stereotypes,
etc., but I don't see the benefit.  Straight vanilla UML intentionally
avoids showing flow of data or control in use case diagrams.

-Eric

[EMAIL PROTECTED] wrote:
> 
> You seem to have lost the plot too.  Actors can be classed as primary or
> secondary actors to a given use-case.
> 
> Primary actors initiate the use case.
> 
> Secondary actors are "used" in some way by the use case in order for it to
> complete its basic or alternate path.  Secondary actors are normally other
> systems or components within a system, although sometimes (I can't think
> when) they could be humans.
> 
> The diagram does not show the flow of information as you say, but show the
> interactions/involvement between use cases and actors.
> 
> Haydn
> 
> 
>                     "Baynes, Steve"
>                     <stephen.baynes@eds        To:
[EMAIL PROTECTED]
>                     .com>                      cc:
>                     Sent by:                   Subject:     RE: (ROSE) Use
case to actor Yes or
>                     owner-rose_forum@ra        No
>                     tional.com
> 
> 
>                     09/10/2001 14:48
>                     Please respond to
>                     "Baynes, Steve"
> 
> 
> 
> This discussion is quite interesting but many people seem to be missing
the
> point.  The use case diagram shows involvement.  All actors are involved
in
> the use case and therefore point to the use case.  The diagram is not
> showing the flow of information.  This is modelled using other artefacts
> (activity diagram, collaboration diagram etc).
> 
> Actors are involved in use cases, use cases are not involved in actors,
any
> actors.
> 
> Hope this helps.
> Steve
> 
> -----Original Message-----
> From: Kelly, Stephen [mailto:[EMAIL PROTECTED]]
> Sent: 09 October 2001 14:04
> To: [EMAIL PROTECTED]
> Subject: RE: (ROSE) Use case to actor Yes or No
> 
> In this example you will clearly end up with an artificial representation
> of
> what you really want to say.
> 
> A real person such as a customer service clerk cannot have something done
> to
> them by the system. In real life, what would happen is the system would
> generate an email (action by the system), and at some arbitray
> (undefinable)
> point the customer service clerk would read emails (action by the person
> actor). The fact that the system generates an email doesn't guarantee that
> the email will be read (the whole customer service department might
quit!).
> Also, setting up a username and password doesn't happen as a result of the
> email being generated. It happens as a result of the customer service
clerk
> doing something to the system.
> The fact that the customer service clerk might decide to do something as a
> result of reading the email is nothing to do with the system - i.e. the
> actions of a person actor cannot be predicted by a use case within a
> system.
> 
> The situation would be far better documented as a 'request new user
> registration' use case, intiated by the broker, which interacts with the
> email system. And then a separate and independent use case like 'create
new
> user' initiated by a customer servive 'actor'.
> 
> The other important thing to remember is that use cases don't really say
> anything about the 'flow of control'. They are supposed to represent the
> functions / actions of a system. So, don't make the mistake of thinking in
> terms of 'this use case happen, then the next use case happens, etc.'.
> 
> Stephen.
> 
> > -----Original Message-----
> > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> > Sent: 09 October 2001 13:12
> > To: Kelly, Stephen; 'Jo�o Paulo Marto Pereira';
> > [EMAIL PROTECTED]
> > Subject: RE: (ROSE) Use case to actor Yes or No
> >
> >
> > Hi,
> >
> > Stephen mentioned that a use case should never point to a
> > person actor.
> > From my point of view, this is not true. The actor who
> > initiates the use
> > case points to the use case. The use case points to actors
> > (persons or systems)
> > who participate in that use case later on.
> >
> > Example: A broker registers himself online in a financial
> > trading system.
> > Customer Service reveives an email to check the registration
> > and provide
> > username and password for the broker.
> > Here the broker initiates the use case wheras customer
> > service enters the
> > game due to the initial request of the broker.
> >
> > I hope that helps.
> >
> > Regards
> > J�rg
> >
> > ___________________________________________________________
> > Joerg Dirbach  mailto:[EMAIL PROTECTED]
> > Software Architect / Software Engineering Trainer
> > Object Technology Center
> > Zuehlke Engineering AG  Wiesenstrasse 10a  CH-8952 Schlieren
> > Phone: +41 (0) 1 733 65 71   Fax: +41 (0) 1 733 69 02
> > http://www.zuehlke.com/
> >
> >
> > -- Original-Nachricht --
> >
> > >
> > >The example you gave is a clear case where it is a very bad
> > idea to relate
> > >a
> > >use case to an actor.
> > >
> > >It makes no sense at all that a 'system' should do something
> > to a 'person'.
> > >Even if you imagine the system flashing a message on that
> > 'person's' screen,
> > >the system is not really doing something to the person - it
> > is simply doing
> > >something to a system device.
> > >
> > >You should never have a causal relationship _from_ a use
> > case _to_ a person
> > >'actor'.
> > >
> > >On the other hand, many analysts treat external systems as
> > 'actor's. In
> > this
> > >case you might have a relationship from one of your use cases to the
> > >'actor'.
> > >
> > >For example, one of your requirements might be that your
> > system send data
> > >to
> > >an external accounting system for the company's book-keeping.
> > >In this case you could represent the accounting system as an
> > 'actor' and
> > >may
> > >have a use case like 'post accounting item' which is related
> > to this actor.
> > >
> > >When doing this it is a good idea to ensure that in your diagrams the
> > >stereotype display of 'actor's that are really external
> > systems is different
> > >to the stereotype display of people 'actor's - use the icon
> > (a stick man)
> > >for people and the label (a box with the <<actor>> label)
> > for external
> > >systems.
> > >
> > >Stephen.
> > >
> > >> -----Original Message-----
> > >> From: Jo�o Paulo Marto Pereira [mailto:[EMAIL PROTECTED]]
> > >> Sent: 09 October 2001 11:42
> > >> To: [EMAIL PROTECTED]
> > >> Subject: (ROSE) Use case to actor Yes or No
> > >>
> > >>
> > >>
> > >> Hi, again
> > >>
> > >> I've had answers saying I could point a use case to an actor,
> > >> an others that
> > >> said, no way! Is it possible and desireble or not?
> > >>
> > >> Thanks
************************************************************************
* Rose Forum is a public venue for ideas and discussions.
* For technical support, visit http://www.rational.com/support
*
* Admin.Subscription Requests: [EMAIL PROTECTED]
* Archive of messages: http://www.rational.com/support/usergroups/rose/rose_forum.jsp
* Other Requests: [EMAIL PROTECTED]
*
* To unsubscribe from the list, please send email
*
* To: [EMAIL PROTECTED]
* Subject:<BLANK>
* Body: unsubscribe rose_forum
*
*************************************************************************

Reply via email to