Book a Demo

Author Topic: Tunneling associations throughout diagram borders  (Read 2280 times)

dirah

  • EA Novice
  • *
  • Posts: 9
  • Karma: +0/-0
    • View Profile
Tunneling associations throughout diagram borders
« on: August 22, 2008, 10:04:29 pm »
Hello,

I would like to show association between software components in a special way. I am depicting a software architecture and use abstracted associations between components (via ports). In this layer I do not show the amount of the signals and the names of the single signals, I just want to show the mere interaction of the components.

When going deeper inside the components at some point I have to show any single signal that goes in or out. This is typically done inside a component diagram which shows the inner structure of any component. When I draw a signal line from a inner component it would be right to choose a target component (as end of the association) that probably is in a different (refined) component diagram. As it is not in the very same daigram I don't get a line inside my diagram.

I solved it so far by drawing a line manually to the outer ports of the component. When coming to a higher component the single signal is now masked inside my abstract interaction (mentioned above). When I refine the compnent containing the real target of the single signal I have to do the same again manually. Firstly draw the absteracted associations down the ports until they are separated into single signals, then I draw a line from the port to the real target showing that very single signal.

This is not only very exhausting, it is also very vulnerable to channges causing inconsistencies. I dream of the following solution: I create a direkt association linking the two components in the different diagrams and the tool searches it way to create the link in the diagrams. It should somehow "tunnel" the signal when coming to the outer border and put it in the abstract signal. In the refinement diagram containing the target it is separated and shown as concrete single signal. This woulg spare manual work and help improve the correctness. When I rearrange the inner structure the new signal ways would fine themselves automatically.

Is there any way to teach EA how to find and draw associations for sch special issues, could this be done in UML anyway or would i need a different language or tool? UML is right for my depictions, but I would like some more automatic features that help the modeling.

Greetings

dirah