[For some reason only today I got this message from Erick Hagstrom (it
was posted on April 24). Stuck in the moderation queue?]

> As for UML 2.1 action semantics, they are EXTREMELY low level
> and flexible. They support state machines and signal-based
> communication, yes. They also support operation invocations and Petri
> net semantics. It's hard to imagine an enterprise IT kind of system
> that couldn't be implemented with all of those tools in the toolbox.

Erick, then how would you go about modeling a query operation that
returns all products ordered by customer 'John Doe' in 2006 using only
UML 2.1?

Thanks,

Rafael

Hagstrom, Erick G. wrote:
> Let's not confuse two different things.
>
> Executable UML is a highly restricted profile of whichever version of
> UML you happen to prefer. The commercially available action languages
> support queries right out of the box, while collections are either
> architectural (not visible to the modeler and not needed) or implemented
> by the modeler as a one-to-many relationship. Either way, it supports
> what people need. My very first Executable UML project was a large scale
> IT kind of system. Worked great.
>
> As for UML 2.1 action semantics, they are EXTREMELY low level
> and flexible. They support state machines and signal-based
> communication, yes. They also support operation invocations and Petri
> net semantics. It's hard to imagine an enterprise IT kind of system that
> couldn't be implemented with all of those tools in the toolbox. Oh, and
> BTW, UML action semantics came from the Executable UML vendors. And OMG
> is currently defining full execution semantics for a small but
> computationally complete subset of UML 2. In other words, no one at OMG
> feels that UML 2.1 limits you to embedded, and they come from a wide
> variety of backgrounds.
>
> But that doesn't really address the original question. Muhammad clearly
> has some problem that is stumping him. What he needs, I think, is some
> specific answer to a specific question. But I haven't really seen the
> question yet. So, Muhammad, can you explain where you're stuck?
>
> P.S. Thanks Ian, cool reference.
>
> ------------------------------------------------------------------------
> *From:* executableuml@yahoogroups.com
> [mailto:executableuml@yahoogroups.com] *On Behalf Of *Rafael Chaves
> *Sent:* Monday, April 23, 2007 12:12 AM
> *To:* executableuml@yahoogroups.com
> *Subject:* Re: [executableuml] Regarding Executable UML in Embedded
> Systems... Require Help
>
> I agree with you, Erick, but I think I can see where Muhammad is coming
> from. Most application examples, success cases and features in
> Executable UML seem to be around embedded systems. Also, if you look for
> instance at the action semantics support in UML 2.1, which is supposed
> to support executable models, it seems it has full built-in support for
> applications based on state machines and signal-based communication, but
> it lacks support for features most required by traditional information
> systems applications, such as queries and collections.
>
> Hagstrom, Erick G. wrote:
> > I'm afraid I'll need to know more about your specific problem. There's
> > nothing in Executable UML that makes it more applicable to embedded
> > systems than elsewhere. You can use Executable UML for any system.
> >
> > So where are you stuck?
> >
> > ----------------------------------------------------------
> > *From:* executableuml@yahoogroups.com
> <mailto:executableuml%40yahoogroups.com>
> > [mailto:executableuml@yahoogroups.com
> <mailto:executableuml%40yahoogroups.com>] *On Behalf Of *MUHAMMAD USMAN
> > *Sent:* Friday, April 20, 2007 10:35 AM
> > *To:* executableuml@yahoogroups.com
> <mailto:executableuml%40yahoogroups.com>
> > *Subject:* [executableuml] Regarding Executable UML in Embedded
> > Systems... Require Help
> >
> > Hello.
> >
> > Am currently stuck at a point where i require some direction to answer
> > the question that why we can't use Executable UML in normaly system? as
> > we know that Executable UML is very useful in Embedded system. Kindly
> > give me some directions regarding how to proceed ahead, any we
> > reference or any view & experiances which u like to share.
> >
> > Regards
> >
> > M.Usman (m_usman99@... <mailto:m_usman99%40yahoo.com>
> <mailto:m_usman99%40yahoo.com>)
> >
>