Before delving into the details of the parallel decomposition approach, it is important to address a common criticism of decomposition: decomposition, it is argued, is an inferior modeling practice because its results are not repeatable. That is, two analysts given the same subject are likely to decompose it in two different ways. As a result, decomposition is considered by its critics to be a slipshod, haphazard approach to subdividing a problem space.
The parallel decomposition approach successfully addresses this issue of non-repeatability. Cross-checking between data and activities, and following guidelines for subdividing by business object type classification and by business object type life-cycle, together with confirmation by event analysis, help to minimize the randomness generally associated with decomposition. Confirmation by event analysis is described in the chapter "Analyzing Activities."
|
Copyright © 2014 CA.
All rights reserved.
|
|