| dc.description.abstract | As the maturity of COSYSMO increases, users continue to identify areas in which the model can be
improved. Recent emphasis has been placed on the clarification of counting rules for the COSYSMO size
drivers. These drivers represent various attributes of the total size of the task of the systems engineering
effort estimated by COSYSMO; in terms of person months. The intent of these rules is to ensure consistent
interpretation and use of the size input parameters that include: requirements, interfaces, algorithms, and
operational scenarios. Experience in applying these rules has exposed a limitation of the current version of
the model; there was no way of including the affect of reusing system components in the calculation of
systems engineering effort.
This has resulted in inaccurate estimates of systems engineering effort for systems that incorporated
significant reuse, as in the case of programs with a high degree of COTS integration. As a result, a method
was needed to account for the fact that not all of the requirements that drive systems engineering effort are
new. Specifically, some of the requirements for a new system may be “reused” from a prior system.
Further, some of the new system’s requirements may be “modified” from a prior system. Moreover, the
evolution of system requirements over the system life cycle may result in “deleted” requirements from the
initial configuration baseline.
On the surface, the notion of reuse in COSYSMO may appear as a necessity-is-the-mother-of-invention
activity but in reality it was an inevitable feature. One reason is that most software cost estimation models
– especially COCOMO II – go into great detail in addressing aspects of software reuse. The other is that
reuse is more prevalent among defense contractors that aim for higher productivity gains as they avoid
pursuing designs from scratch.
For these reasons, this paper provides (1) an approach for handling reuse in systems engineering in terms of
the number of systems requirements in COSYSMO, (2) a discussion on the potential cost drivers that could
be influenced by reuse, and (3) strategies in which this approach can be extended to include the three other
size drivers in the model. | en_US |
