作者:
A Newman
关键词:
摘要:
ABSTRACT How graph theory informs the design and maintenance of large scale software systems. Abstract: The networks of interdependencies in large-scale software systems are complex, difficult for designers and developers to understand, and are a significant driver of the costs to develop and maintain software systems. Interdependencies and the costs they drive result directly from design structures and yet a disproportionately small effort is allocated to the design phase within the lifecycle of a product. We propose that object-oriented (OO) design structures and their costs can be better understood via the application of graph-theoretic representations and metrics. In the present work, we develop a general method for representing OO designs as graphs, we develop a set of metrics for identifying problematic dependency structures and for quantifying their effect on defect repair costs, and we present the results of three empirical studies applying the metrics to a set of software applications.
在线下载