64th ISI World Statistics Congress - Ottawa, Canada

64th ISI World Statistics Congress - Ottawa, Canada

Validating an open-source software library


Ryan Lekivetz


  • J
    Joseph Morgan


64th ISI World Statistics Congress - Ottawa, Canada

Format: IPS Abstract

Keywords: design, experimental-design, validation

Session: IPS 316 - Design and Analysis of Experiments for Data Science

Wednesday 19 July 2 p.m. - 3:40 p.m. (Canada/Eastern)


Validating complex software systems is an extraordinarily difficult task. Test engineers are faced with the challenging prospect of ensuring that a software system satisfies its requirements while working within a strict budget. Algorithms in an open-source library are often developed independently by several different development groups and then integrated into the library. Although we may assume that the individual algorithms have been rigorously tested by individual development groups, it may be unreasonable to expect that validation of the integrated library is as rigorous.
Choosing a test suite for such an endeavor can be framed as a design of experiments problem. Combinatorial testing is a software testing methodology that may be viewed as a design of experiments approach to addressing the software testing challenge. We provide an example, using an open-source software library, to illustrate some of the challenges with validating such a system.