Academic
Publications
Clustering Software Artifacts Based on Frequent Common Changes

Clustering Software Artifacts Based on Frequent Common Changes,10.1109/WPC.2005.12,Dirk Beyer,Andreas Noack

Clustering Software Artifacts Based on Frequent Common Changes   (Citations: 44)
BibTex | RIS | RefWorks Download
Changes of software systems are less expensive and less error-prone if they affect only one subsystem. Thus, clusters of artifacts that are frequently changed together are subsystem candidates. We introduce a two-step method for identifying such clusters. First, a model of common changes of software artifacts, called co-change graph, is extracted from the version control repository of the software system. Second, a layout of the co-change graph is computed that reveals clusters of frequently co-changed artifacts. We derive requirements for such layouts, and introduce an energy model for producing layouts that fulfill these requirements. We evaluate the method by applying it to three example systems, and comparing the resulting layouts to authoritative decompositions.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...Clustering criteria for layoutbased clustering have been formally and thoroughly investigated by Noack [23], and the method was applied to software graphs in CCVISU [7, 10], which is the technological basis of FEATUREVISU...
    • ...FEATUREVISU uses Noack’s clustering energy model, which has been successfully applied to layout-based clustering in a number of different domains [6, 7, 10, 23]:...
    • ...Visual clustering and CCVISU have been used previously to decompose software graphs into subsystems [6, 10]...
    • ...It has been shown that visual clustering can aid program comprehension by visualizing the software design based on distances in the clustering layout [6, 10, 23]...
    • ...Layout-based clustering can aid program comprehension [6, 10, 23]...

    Sven Apelet al. Feature cohesion in software product lines: An exploratory study

    • ...Beyer and Noack (2005) introduced a clustering method, using co-change graphs, to identify clusters of artifacts that are frequently changed together...

    Maen HammadMichaelet al. Automatically identifying changes that impact code-to-design traceabil...

    • ...Further, transactions have been used to: identify sub-systems that fulfill the requirement of independent evolution [19] find strong inter-module dependencies [15], [20] identify sets of co-changed files [21], [22] relate software entities and non-program artifacts and find traceability links [1], [23], [24] identify cross cutting concern candidates [25]...

    Adam Vanyaet al. Approximating Change Sets at Philips Healthcare: A Case Study

    • ...[11], [12]). In [13], Briand, Daly and W¨ ust identified and organized more than thirty coupling metrics...
    • ...We see a fundamental problem in all this research: the difficulty to validate empirically the validity of the quality metrics (a fact alluded to in [3], [6], [1], [11] for example)...

    Nicolas Anquetilet al. Legacy Software Restructuring: Analyzing a Concrete Case

    • ...Siliarity metrics have included latent semantic analysis of source code [8], software evolution and change [1], and features gathered from static and dynamic analysis of the code [4, 7]. In [9], the authors posit that the quality of clustering results depends heavily on the characteristics of the software to be clustered, suggesting that constraining KADRE’s clustering process to a specific software domain will improve quality...

    David Woollardet al. Kadre: domain-specific architectural recovery for scientific software ...

Sort by: