Re: software architecture/design question

Responding to Ragu...

Thanks Lahman, for quick response clarifying some of the implications
of the application of the mediator. But not sure whether I understood
everything you have told here. I shall study ur response further.

When deciding on mediator pattern here, I speculated that (1) the
sequence may be arbirary, decided at runtime, (2) the set of processors
(A..G) can vary. And I agree that there may need to be more complex
processing than just configuration, I am not much familiar with problem
domain either. (and thus possible bad speculations)

If (1) and (2) are true _during the execution and based on changing execution context_, then I would tend to agree with you. That would entail a dynamic component that would be awkward to deal with for the R1 and R2 relationships in my model. One would have to provide dynamic infrastructure to modify both relationships "on the fly". Mediator happens to provide a convenient infrastructure for that.

However, if (1) and (2) are only true from one entire execution to another (e.g., they are defined and fixed differently at startup), then the R1 and R2 relationships can be instantiated in a static manner based on configuration data without Mediator by using the sort of factory infrastructure one would already have in place to create the [SlotCard] objects.


*************
There is nothing wrong with me that could
not be cured by a capful of Drano.

H. S. Lahman
hsl@xxxxxxxxxxxxxxxxx
Pathfinder Solutions -- Put MDA to Work
http://www.pathfindermda.com
blog: http://pathfinderpeople.blogs.com/hslahman
(888)OOA-PATH



.