[news.eclipse.modeling.mdt.uml2] Re: seperating stereotypes in profiles

["Tas Frangoullides" <tfrangoullides@xxxxxxxxxxxxxxxxxx>]

Hi Robert,

I tend to find that the solution to such problems surface from examining the 
underlying domain and the concepts represented. Somewhere within them will 
be a natural boundary you can use. The question I would ask is why is is not 
appropriate for model A to apply stereotype B. Is this just a case for this 
partciular instance of model A or for all type A models?

It sounds like you have common elements to both types of model. To avoid 
duplication you could define 3 profiles, 1 containing the common stereotypes 
which can be used by either type of model, and then 2 model specific 
profiles which extend the common profile. This will avoid duplication.

Also, there might be a way to add validation to your profiles to ensure they 
follow your rules. I know there is OCL constraint support in models, does 
anyone know if OCL is supported in Profiles?

Tas



"robert vrooland" <vrooland@xxxxxxxxxxx> wrote in message 
news:3bebecb56e6496444d76652fc0d18c40$1@xxxxxxxxxxxxxxxxxx
> Hello.
> I am working on a project where i have to create variations of a certain 
> model.
> I will apply a profile to a model. The profile I create has all the 
> stereotypes. The problem is that the models do not use all the stereotypes 
> in the applied profile.
>
> Say model A applies stereotype 1 and 2 and model B applies stereotypes 2 
> and 3.
>
> Is there a way to make sure that model A can never apply stereotype 3.
>
> The 2 choices I got for this project is to either make a seperate profile 
> for each variation of the needed model. Or to have one profile where the 
> stereotypes are grouped into the variations for the model.
>
> Id prefer using a single profile. But if the profile contains many 
> stereotypes and the model only needs number 1 and 2, it becomes confusing.
>
> tnx in advance for any help.
>
> Robert