Evaluating OSS Projects

Wray and Mathieu’s (2008) article, Evaluating the Performance of Open Source Software Projects Using Data Envelopment Analysis, parallels my research. The study provided excellent insights onefficiency of open source software (OSS). Wray and Mathieu (2008) limited their study to only security based OSS projects, which consisted of 34 projects. The study also utilized a limited number of inputs and outputs with the data envelopment analysis (DEA) model. In my research I hope to further their study by encompassing more projects along with more inputs and outputs in my DEA model.

DEA is a nonparametric linear programming formulation technique that accounts for multiple inputs and outputs to measure the efficiency of decision making units (DMU). In this study and my research, a DMU represents an OSS project. In essence, the DEA will rank DMU in order ofefficiency. From the results, we will be able to discover the indicators of successful efficient software development projects.

Wray and Mathieu (2008) selected OSS developers and bug submitters from 34 security projects on SourceForge.net as their input variables for the DEA. For their output measures, the study focused on the number of software downloads and the SourceForge.net project rank. SourceForge.net is the world’s largest OSS project hosting database that houses over 100,000 OSS projects with over 1 million registered users. SourceForge.net ranks projects weekly based upon traffic, communication, and development statistics collected on each project.

Wray and Mathieu (2008) discovered that the most relative efficient projects when compared to the others were Ophcrack, ClamWin FreeAntivirus, Simple Python Keylogger for Windows, ShellTer, Another File Integrity Checker, Network Security Toolkit, J2EE Certificate Authority, and BlockSSHD. The research conclusions can be used by security based OSS project managers to determine the relative efficiency of their project against other similar projects. The study can help with critical decisions on work effort allotment to project areas and assignments to produce larger yields and benefits.

Reference

Wray, B., & Mathieu, R. (2008). Evaluating the performance of open source software projects using data envelopment analysis. Information Management & Computer Security, 16(5), 449-462.