Re: MSC-News: Problem with instance-oriented - shared in-line expression textual


Subject: Re: MSC-News: Problem with instance-oriented - shared in-line expression textual
From: Ekkart Rudolph (rudolphe#informatik.tu-muenchen.de)
Date: Tue Oct 06 1998 - 13:36:39 GMT


The originator of this message is responsible for its content.
-----From Ekkart Rudolph <rudolphe#informatik.tu-muenchen.de> to mscnews -----

Dear Paul,
first of all I think the syntax description needs slight correctons:

instance i;
        alt begin shared j;
                action a; /* is this event part of the same choice as action b? */
         alt end;
endinstance i;
instance j;
         alt begin shared i;
                 action b; /* is this event part of the same choice as action a? */
         alt end;
 endinstance j;

I can answer both questions in the comments with yes. Event a is part of the same
choice as b. In fact, you have only one choice (one alternative) in your example.
Because of the shared construct, the event a on instance i and event b on instance j
belong to the same alternative. Of course, a formal argument can be given only
by means of the formal semantics contained in Annex B. For that I refer you to the
semantics department (Sjouke Mauw).

Best regards!
Ekkart
=================================================================================
Dr. Ekkart Rudolph phone: +49 89 289 25361
Technische Universitaet Muenchen fax: +49 89 289 28183
Institut fuer Informatik, email: rudolphe#informatik.tu-muenchen.de
c/o Lehrstuhl Prof. Broy
Arcisstrasse 21
D - 80290 Muenchen
=================================================================================

>
> Hi Ekkart,
>
> > Dear Paul,
> > can you give a simple example? I cannot see immediately your point because
> > of the shared construct.
> >
>
> If we have two action boxes a and b each of which reside upon different
> instances (I, J) within an 'alt' inline expression, it is not possible
> to determine whether the expression has the choice of executing:
>
> - action a followed by action b, or vice-versa
>
> or
>
> - action a or action b, not both.
>
> Using the above example the instance-oriented grammar would look as
> follows:
>
> instance I;
> alt begin i,j;
> action a; /* is this event part of the same choice as action b? */
> alt end;
> endinstance I;
> instance J;
> alt begin i,j;
> action b; /* is this event part of the same choice as action a? */
> alt end;
> end instance J;
>
> I hope this helps.
>
> Kind Regards
>
> Paul.
> --
>
> ====================================================================
> | Paul Baker | |
> | | |
> | European Research Laboratory | SMTP: paul#erl.comm.mot.com |
> | Land Mobile Product Sector. | MSMail: Baker Paul |
> | Motorola Ltd | E-mail: CPB005 (X.400) |
> | Basingstoke, England (UK) | Voice: +44 1256 48-4117 |
> | Maildrop: EUR-UK, location ZUK02 | Fax: +44 1256 484490 |
> ====================================================================
>
>
> Ekkart Rudolph wrote:
> >
> > The originator of this message is responsible for its content.
> > -----From Ekkart Rudolph <rudolphe#informatik.tu-muenchen.de> to mscnews -----
> >
> > Paul Baker wrote:
> >
> > > Dear MSC folks,
> > >
> > > I think there is a problem regarding the instance-oriented textual
> > > grammar for the shared
> > > inline ‘alt’ and ‘par’ expressions (section 5.3 of Z.120), which:
> > >
> > >>
> >
> > > (2) makes it impossible to define the semantics for alt and par
> > > expressions using the
> > > instance-oriented grammar. The reason for this is that it is not
> > > possible (in some cases)
> > > given the information provided to determine which events, existing on
> > > different instances,
> > > are associated with different alt or par arguments. And in fact
> > > whether events are
> > > indeed associated with the same alt or par expression.
> >
> > Dear Paul,
> > can you give a simple example? I cannot see immediately your point because
> > of the shared construct.
> >
> > Apart from that, I do not like so much the terminology 'instance oriented' and
> > 'event oriented' grammar in MSC-96. In fact, my idea was to include the
> > 'instance oriented' representation of MSC-92 within the new 'event oriented' representation.
> > (since message inputs may appear before corresponding message outputs within this
> > 'event oriented' grammar)
> >
> > Unfortunately, because of upward compatability there still remained this
> > <old instance head statement>. Apart from that, I would rather distinguish between
> > <shared event> - and <multi instance event> representation (e.g. with respect to
> > conditions and inline expressions). I must admit that the <shared event> - representation
> > in some parts (e.g. inline expression) resembles the original 'instance oriented'
> > representation, in order to avoid the obscuring mixing of events on different instances
> > within operator expressions.
> >
> > Best regards!
> > Ekkart
> >
> > =================================================================================
> > Dr. Ekkart Rudolph phone: +49 89 289 25361
> > Technische Universitaet Muenchen fax: +49 89 289 28183
> > Institut fuer Informatik, email: rudolphe#informatik.tu-muenchen.de
> > c/o Lehrstuhl Prof. Broy
> > Arcisstrasse 21
> > D - 80290 Muenchen
> > =================================================================================
> >
> > -----End text from Ekkart Rudolph <rudolphe#informatik.tu-muenchen.de> to mscnews -----
> > For help, email "majordomo#sdl-forum.org" with the body of your email as:
> > help
> > or (iff this does not answer your question) email: owner-mscnews#sdl-forum.org
>

-----End text from Ekkart Rudolph <rudolphe#informatik.tu-muenchen.de> to mscnews -----
For help, email "majordomo#sdl-forum.org" with the body of your email as:
    help
or (iff this does not answer your question) email: owner-mscnews#sdl-forum.org



This archive was generated by hypermail 2a23 : Wed Jun 19 2013 - 13:16:38 GMT