Title of the Paper Applied Software Development Risk Model

The Process of developing large software systems creates many source code files with complex inter-dependencies. Clusters of mutually dependent files introduce the possibility of a chain of forced consequential changes when a single cluster member file is changed. The software development risk model, developed here, shows that the density of dependencies within such clusters plays a crucial role in this behavior. We develop a file-rank procedure which orders the entire system’s file set by increasing risk. This ranking process should prove to be useful while managing the development of large systems, indicating where attention should be focused to reduce development risk. We have applied this model to code from several projects with interesting results.