Paper Title
A Slicing-Based Effort Estimation Approach For Open-Source Software Projects
Abstract
A slicing-based approach to estimate maintenance effort for open-source software projects is presented. GNU
Linux kernel is used as a case study with over 900 versions spanning 17 years of history. For each version the forward static
slice of all variables is computed using a lightweight slicing approach. The slice size is computed as the total number of
lines, functions, or files in the slices. Changes to the system are then modeled using the difference between the slice sizes of
two versions. The hypothesis is that this model is predictive of maintenance effort. The three different granularities of slices
(i.e., line, function, and file) are analyzed. The results demonstrate that the approach accurately predicts effort in a scalable
manner.
Keywords- effort estimation; program slicing; maintenance