Rose Question, automatically create dependencies/associations from Sequence Diagram

[Richard Klein]

I have a question about Rational Rose.
Suppose I create a class diagram and a sequence diagram. In the sequence diagram I send messages to objects. Is it possible to automatically create the dependencies/associaties in the class diagram?
Regards
Richard

[noel angel]

Hi Rich,
If you have added functions to your class diagrams you may select those messages for a sequence diagram but at this time Rose will not populate the classes with your message names automatically. There is a report & checker in Rose that will check your sequence diagrams to make sure all of the messages exist in some class somewhere but it doesn't populate the classes.
Consider the fact that if you name the functions slightly differently in your sequence diagram from the class they will fall out of the tool as an exception or problem if you like that word better. Even case differences in the name will cause this name to appear in the report. Then you can fix them and make all of your names consistent. This can be a real pain!
I hope that I have answered your question.
Noel
SCJP, SCWCD, SCEA(Part 1)

[Richard Klein]

Hi Noel
Thanks for your repsonse
Regards
Richard

[Juan Rolando Prieur-Reza]

Originally posted by noel angel:
...at this time Rose will not populate the classes with your message names automatically...

There is a good reason for not automating the update of Class diagram from
the method signatures and other parts of the class. The source of a dependency can be (1) a method arg, (2) reference to a "globally" accessible object, such as a static, or (3) other reasons.
The first problem is that you could easily clutter the diagram with numerous dependency lines. Another problem is that "important" dependencies that are not necessarilly identifiable as source (1) or (2) above.
However, it would be nice to have a button that would populate all findable dependencies by (1) or (2), so you can then hide whatever you don't want to show. Something like that. But to maintain round-trip engineering functionality of the diagramming tool, what should you do when a user just manually adds and names a dependency arrow that may or may not map to situation (1) or (2)? I don't know.
[ November 15, 2003: Message edited by: john prieur ]