Refactored VaVe metamodel

[SofiaAnanieva]

This is a draft pull request for the refactored VaVe metamodel. It is the first step of VaVe refactorings (branch variability). Further changes may follow.

[JanWittler]

I had a first look at the model, didn't find anything problematic or surprising to me. Some minor comments:

• relations should follow camelCaseNotationStyle (e.g. rootFeature instead of rootfeature)
• there are multiple to-many relationships with a singular name, should be plural
• why are Feature.treeconstraint and TreeConstraint.feature not setup as opposites?

Besides that we obviously need documentation but as the PR is still a draft, I guess this will come in the progress. We should also adjust the main Readme to document the new variability top level element in the Project Structure section.

[SofiaAnanieva]

Thanks a lot for your helpful comments! I just added a commit to this draft pull request to address them. Regarding your third point about why Feature.treeconstraint and TreeConstraint.feature are not opposites: That was because the former represented the child tree constraints of a feature and the letter represented the child features of a tree constraint, and neither represents the parents (i.e., neither of the two were bi-directional relations). I addressed this now by renaming the references to TreeConstraint.childFeatures and Feature.childTreeConstraints and giving them parent-opposites to make them bi-directional.

[JanWittler]

So if I want to iterate over some feature tree I'd have to go childFeatures -> childTreeConstraints -> childFeatures and so on (i.e. alternate between Feature and TreeConstraint) ? This seems to me as if we should extract this into a proper tree data structure.

[JanWittler]

Closed due to inactivity