Master Thesis Defense: Christopher Forbes

Speaker: Christopher Forbes

Supervisor: Dr. J. Rilling

Examining Committee: Drs. G. Butler, J. Paquet, J. Yang (Chair)

Title: Oswaldo: A Semantic Web Enabled Approach for Identifying Open Source License Violations

Date: Wednesday, August 29, 2018

Time: 10:00 a.m.

Place: EV 11.119

ABSTRACT

Open source license violations are numerous, multifaceted, and pose significant risk to developers and companies in the form of litigation, sometimes resulting in millions in dollars in damages or settlements. Free/Libre and Open Source Licenses utilize copyright law and are written in legalese, which is often outside the scope of a developer’s expertise. Software Engineers commit violations of these licenses’ terms and conditions easily and often unknowingly. Consequently, increased knowledge, better tools, and sound processes to detect and prevent license violations are extremely important. This work is an investigation in the types of potential license violations that are committed, through direct and transitive dependency hierarchies in hundreds of thousands of real-world software projects. This thesis contributes a novel approach, entitled Oswaldo, that defines and detects three types of license conflicts: Type 1 Simple Violation, Type 2 Embedded Violations, Type 3 Compound Violations. Unidirectional compatibility/incompatibility relationships of major licenses are modelled. Ontologies and Linked Data are advantageously exploited to detect transitive violation Types 2 and 3, as well as the direct violation Type 1. This thesis also reports initial evaluations of these three types of license violations found in the Maven repository.