The
Analysis model contains the description of what the application is supposed
to do and the Design model describes how it is going to be done. In an
iterative process you may have to update several times the previous content
of both models. An analysis class (whatever its stereotype is) may evolve
into 1 or several design classes, whether they are boundary, control or entity.
I tend to keep both models in my logical views along with their respective
sub-decompositions in layers and/or packages and/or subsystems so that there is
a trace of what was the initial understanding. At the beginning of the iterative
process the design classes are basically duplicate of the analysis classes and
it is changing over time.
JPM
-----Original Message-----
From: Keysers, Wonne [mailto:[EMAIL PROTECTED]]
Sent: Friday, January 25, 2002 12:03 AM
To: '[EMAIL PROTECTED]'
Subject: (ROSE) Analysis -> designHi,
When opening the RUP model which comes with rose, there is an
Analysis Model and Design Model package.
Question: what should reside in each of those packages?During analysis, resulting objects are still at rather high level of abstraction. (boundary, entity or control) When going from analysis, do we re-use the analysis classes or do we create new ones?
The second option is the easiest one, but then the analysis diagrams, which are a nice documentation of the overall functionality of the system, are lost.
Anyone?
Thanks
Wonne
