INRIA Associate Team

Source de subvention

European Smalltalk User Group

Institut National de Recherche en Informatique et Automatique

Professeur(e)s impliqués

Résumé

Software is everywhere. Software is running the world and the world is always changing. Software evolution is inevitable. Software evolution is difficult because the life expectancy of large software systems can be more than 50 years! Conversely, it is also difficult because of the recent systems of systems that integrate physical devices in Internets of Things. When dealing with such systems, software engineers spend most of their time understanding their reasoning and decisions, i.e., their architecture, design, and implementation. The proposed associate team considers three research directions to support the software engineers maintaining and evolving such large software systems: (RD1) system analyses and (RD2) debugging for (RD3) program comprehension.

Expected outcomes:

  • RD1: We plan one journal article on the comparison of Moose and Ptidej and their characteristics. We plan another journal article on a practical evaluation of the integration of Moose and Ptidej. We will also release open-source bridges integrating Moose and Ptidej as well as related tools (e.g., JavaParser). The work on algorithm analysis will be the topic of at least one conference paper. A journal version will be realized as part of RD3.
  • RD2: We plan one journal article describing and evaluating the language to build debuggers. We also plan a series of empirical evaluations of Swarm Debugging with our built-for-purpose debuggers. We will release as open-source the language and all produced debuggers, and in particular, a Swarm Debugging plugin for Pharo.
  • RD3: We plan to develop a set of customizable views that will be integrated into Moose and Ptidej and released as open-source. This work and the meta-data needed to adapt/combine views will lead to at least one conference paper. It will also be the topic of a journal paper for its use of the systems analyses developed in RD1. The evaluation of the views and their (dis)advantages to software engineers will be the topic of one journal paper.