Applying a Rule-Based Natural Language Classifier to Open Source Requirements: a Demonstration of Theory Exploration

Open source projects requirements are mostly informal, text descriptions found in requests, forums, and other correspondence. Understanding of such requirements can provide insight into the nature of open source projects. Previously, we have demonstrated the Requirements Classifier for Natural Language (RCNL), which aids in NL requirements analysis. Herein, we demonstrate how the RCNL can aid in theory building. From its application to 16 open source projects, we conjecture a simple wave theory of requirements innovation: innovations expressed in requirements appear as a wave that is reflected in a subsequent wave of features that is reflected in a subsequent wave of product downloads. Although the theory is a conjecture, the process of its exploration demonstrates how RCNL can be used to explore theories about open source projects -- theory exploration that would otherwise be intractable because of the difficulty in analyzing NL artifacts for requirements properties.