#639 Modeling top level interface - #714
Conversation
- `PokaYokeProfileServices`: add `isInterface` and `setInterface` - `PokaYokeUmlProfileUtil`: add `isInterface`, `setInterface`, `isFormalActivity`
| } | ||
| } | ||
|
|
||
| private void checkValidInterfaceActivity(Activity activity) { |
There was a problem hiding this comment.
- Should we disallow templated/parameterized interface activities, or not?
- Should we disallow interval constraints in interface activities, or not?
- Should we disallow pre/post conditions in interface activities, or not?
There was a problem hiding this comment.
I would say no parameterized interface activities and no interval constraints in them; I would allow pre/post-conditions. What do you think?
There was a problem hiding this comment.
I was thinking the same regarding no parameters and no interval constraints. I was also thinking to allow precondition. I'm not sure about a post-condition. Would an interface not go on forever, generally?
There was a problem hiding this comment.
I think the main use we want to have is indeed to have an infinite loop, but it doesn't have to be the case; we can allow a "simple" interface. Maybe we can add an assertion that checks if there is a final node then we can have a postcondition, otherwise not?
There was a problem hiding this comment.
I assume you don't mean a code assertion that makes validation crash, but rather a mean model validation constraint? I like it. It does allow having postconditions, but only when it makes sense to have them.
There was a problem hiding this comment.
Re: parameterized activities.
I cannot find a method similar to isParameterizedActivity. There are methods in the CIF context asking if it contains any parameterized activity. Should we add a dedicated method? Probably new issue as well.
There was a problem hiding this comment.
Sure, new issue is fine.
Co-authored-by: Dennis Hendriks <dh_tue@hotmail.com>
|
Sorry, I realized now that one commit was supposed to be fixed-up. I had some issues with the interactive rebase, I thought it went through -- but clearly it didn't. |
|
Last commit has wrong issue number prefix. |
2 participants
Addresses #639, just the modeling part. Best review commit-by-commit.