Subject: RE: Connect construct
From: Birger Møller-Pedersen (ETO) (Birger.Moller-Pedersen#eto.ericsson.se)
Date: Tue Jun 05 2001 - 08:07:34 GMT
Dear Rick,
this is an interesting generalization. Apparently we only went half the way in regarding the unnamed exit (corresponding to final state and default exit transition in statecharts) as just a special case of exit. Your proposals seem right, so the only thing we should consider if there is a substantial difference between "finalizing" a composite state and just exiting it.
/birger
> -----Original Message-----
> From: Rick Reed [mailto:Rick.Reed#etsi.fr]
> Sent: 4. juni 2001 10:22
> To: 'meeting#sdl-forum.org'
> Cc: stf188
> Subject: Connect construct
>
>
> Dear All,
>
> Observation resulting from ETSI STF188 work.
>
> A <connect area> (or in textual grammar a <connect part>)
> represents the
> transition to be taken on exit from a composite state. It contains an
> optional <connect list>, which defines the exits that lead to the
> transition.
>
> The <connect list> can be omitted, to denote and then denotes
> the unnamed
> exit from the composite state, or it can be a list of
> <\underline{state exit
> point} name>s or an <asterisk>.
>
> If the composite state has an unnamed exit, and named exits
> (for example, an
> unnamed return, and returns named A and B). It is not
> possible to include
> the unnamed return and a subset of the other returns in the
> <connect list>.
> If in the example, the same state is to be taken for the
> unnamed exit and
> exit A (but not exit B), these would have to be shown as two <connect
> area>s.
>
> Moreover the Model for <asterisk connect list> is such that
> in the example
> given above, it denotes A and B, but not the unnamed exit.
>
> The abstract grammar for a Connect-node appears to be defined
> in the wrong
> place: it is part of a Composite-state-type-definition, but includes a
> Transition which should lead to a Terminator or Decision-node
> in the context
> in which the Composite-state-type-definition is used. In my opinion
> Connect-node-set should therefore by part of a State-node,
> rather than part
> of Composite-state-type-definition.
>
> --
>
> Suggestions:
>
> 1. The keyword DEFAULT should be allowed in the <connect
> list> meaning the
> unnamed exit (similar to the DEFAULT entry point).
>
> 2. The asterisk notation should also denote the DEFAULT exit.
>
> 3. There should be a condition on the abstract grammar that the
> State-exit-point-name should be defined in the Composite-state-type
> identified by the Composite-state-type-identifier of the State-node.
>
> 4. Change the point in the abstract grammar where
> Connect-node-set appears
> to be State-node.
>
> 5. Allow <asterisk> to be followed by a list of <\underline{state exit
> point} name>s (or DEFAULT) meaning all exits except those listed.
>
> --
> Rick Reed (at ETSI)
> Tel: +33 4 92 94 42 68
> mob: +44 79 70 50 96 50
> email: rickreed#tseng.co.uk
>
This archive was generated by hypermail 2a23 : Mon May 05 2008 - 20:30:55 GMT