Soutenance de thèse Eric Harth
M. Eric Harth soutiendra, en vue de l'obtention du grade de docteur ès sciences mention informatique de la Faculté des sciences de l'Université de Genève, sa thèse intitulée:
Program Understanding - The Narrative Hypothesis
Date: Lundi 21 janvier 2019 à 15h00
Lieu: HEG / Battelle bâtiment B, salle B2.19 (2ème étage)
Program understanding, a central part of software development and maintenance, has been widely studied over the last few decades, and many models have been proposed to describe how engineers comprehend software and what techniques may improve and reduce the time spent in understanding software. However, instead of modelling what understanding is, most of these studies focus on strategies and techniques applied by engineers when trying to understand software. The paradox is that the goal of these techniques is usually ill-defined. How then can we assess their relevance?
The purpose of this thesis is, first, to propose an operational definition of software understanding in the context of software development. Then, we developed a new model of software understanding that is adapted from the strong story hypothesis of MIT’s professor Patrick Winston. This hypothesis suggests that understanding may be equivalent to story manipulation operations through replications, merges or deletion. Based on our model, we show that using NLP and information retrieval techniques (IR), we can “map” the domain-level story (the functionality of the software) to the technical story (i.e., source code), which is associated with the runtime execution of the software. From that background we present the tools that we developed and apply them in case studies.