The Sept 12th meeting of Baltimore SIGAda will feature William R. Bitman of the Johns Hopkins University Applied Physics Lab, speaking on "Reuse & Maintenance Costs".

Many software engineers claim that component design produces software with higher quality, decreased maintenance cost and increased reuse. This talk discusses a technique to isolated design as a factor of maintenance cost.

The Interfacing Techniques Comparison Graph visually compares applications in terms of attributes that relate to maintenance cost. Applications that have both lower coupling and lower complexity lie closer to the origin of the graph and exhibit lower maintenance cost than those that do not. This study supports the idea that compositional techniques are important for achieving these improved metrics.

The graph can be used in three ways. First it serves as a decision support tool for managers to determine whether expected maintenance savings compensate for the additional training, effort and time needed to support compositional development. Second, it functions as a decision support tool for designers and coders as they determine, for each module interface, whether to use coupled techniques or composition. The graph can help identify those situations in which the long term cost gain justifies the extra time needed for compositional design. Third, it can serve as a maintenance cost estimation tool. This study found a close correlation between predicted and actual maintenance effort.

Updated Aug 29, 2000