A modified Yule process to model the evolution of some object-oriented system properties

Understanding and simulating software evolution

“…In their study, logical change size is defined as the commit size as measured by the number of files changed, while structure changes are defined as the differences in the file dependency graphs obtained by comparing daily snapshots. Hattori and Lanza (2008) confirmed that commit size, measured by the number of changed files, closely fol- Power law also exits in many other measures of software systems, such as the number of methods, fields, constructors, subclasses, instance variable names and method names (Wheeldon and Counsell, 2003;Concas et al, 2007;Turnu et al, 2011), as well as the number of bugs in a module .…”

“…Class graphs, method call graphs, and library and package dependency networks in software systems are all scale-free networks, i.e., their degree distributions follow a power law (Valverde and Solé, 2003;Louridas et al, 2008;Turnu et al, 2011).…”

“…In this chapter, we construct a software evolution model that uses preferential attachment to simulate the constructive process and SOC to simulate the destructive process. Although preferential attachment and SOC have been used separately to simulate software evolution (e.g., Gorshenev and Pis'mak, 2004;Turnu et al, 2011), having preferential attachment and SOC in one model was never attempted before, to the best of our knowledge.…”

Explore the Community Structure of the Software Network

“…Many software systems have reached such a huge dimension that it looks sensible to treat them as complex networks [1][2], [14]. In particular, software networks made of thousands of nodes show properties typical of complex networks [3], like a power law distribution of the node degree, scale free properties [4], fractal and self-similar features [26], the small world property, power law scaling for bug distribution [15], for refactored classes [16], and so on.…”

Pluggable Controllers and Nano-Patterns in Java with Lola

“…The reuse index is more robust than other statistics such as mean and median [22]. In fact, reuse is characterized by a long tail distribution as it happens in many software engineering contexts [23][24][25][26]. Being based on the h-index, also the r-index presents the same advantages, apart from the mentioned robustness, such as the fact that it captures the "core" of projects' methods, namely those that are more reused.…”