| Re: [henshin-dev] Last-minute model change suggestion |
I'd prefer the semenatics Christian mentioned. I am sure that this behaviour could be implemented in a number of ways but, however, it is simple and straightforward to use ConditionalUnits. Regards, Stefan Am 03.02.2012 13:48, schrieb Gregor Bonifer: > No. The default value is false. This is overriden by the result of the > then- or else-unit. if there is no such unit the default value won't > be overriden and therefore the whole unit returns false. The only > difference between interpreting the then- and else-case is that there > is an additional check if an else-unit is present. I would expect the > same behavior from the then-branch, if it could be empty. > I'm not sure if it is reasonable to allow empty branches in > ConditionalUnits at all. I would introduce a new Unit NegatedUnit, > that holds exactly one subUnit and negates its result. This would > allow to produce the behavior of a ConditionalUnit with empty > then-unit by negating the first subUnit in a strict SequencialUnit. I > would prefer to increment the else-units lowerBound, so > ConditionalUnits can only be used to model real/existing alternatives. > This would eliminate semantic redundancy: A strict/no-rollback > SequentialUnit s with two subUnits a and b is equivalent to a > ConditionalUnit holding the a as if-unit and b as then-unit. Wrapping > a in s with a NegationUnit would correspond to a ConditionalUnit with > a as if-unit and b as else-unit. > > Regards, > Gregor > > Am 03.02.2012 12:44, schrieb Christian Krause: >> I have no objection to changing the multiplicity for the then to >> 0..1. I think it should be consistent with the else part. >> >> What is the semantics implemented in the interpreter if one of them >> is null? I would guess that the unit simply succeeds without >> performing any action. Correct? >> >> Cheers, >> Christian >> >> On 02/03/2012 11:09 AM, Riegerf@xxxxxxxxxxxxxxxxxxxxxxx wrote: >>> Hi everyone, >>> >>> just a quick question/suggestion for a last-minute model change: >>> Currently, the ConditionalUnit's then requires a subunit, whereas >>> the else can have 0..1 subunits. Can we change this so both then and >>> else can have 0..1 subunits? This would require only very minimal >>> code modification and greatly increase expressivity. >>> >>> Felix >>> >>> _______________________________________________ >>> henshin-dev mailing list >>> henshin-dev@xxxxxxxxxxx >>> https://dev.eclipse.org/mailman/listinfo/henshin-dev >> >> _______________________________________________ >> henshin-dev mailing list >> henshin-dev@xxxxxxxxxxx >> https://dev.eclipse.org/mailman/listinfo/henshin-dev >> > > _______________________________________________ > henshin-dev mailing list > henshin-dev@xxxxxxxxxxx > https://dev.eclipse.org/mailman/listinfo/henshin-dev
Attachment:
signature.asc
Description: OpenPGP digital signature