Category: Publications

Mes publications

Actes de la conférence Autour du Libre 2002

Vieux maut tard que jamais 😉

J’ai enfin pris le temps de finaliser la mise en ligne des actes de la conférence Autour du Libre 2002 qui s’était déroulée à l’INT.

Certaines confs sont en ligne avec vidéos (en format SMIL, avec un encodage plus ou moins libre ), ou avec slides.

Bien qu’elles datent un peu, toutes sont probablement toujours d’actualité !

N’oubliez pas, l’édition de cette année est prévue début avril à Brest.

Update : désolé, les vidéos et les slides mis en ligne hier étaient inaccessibles de l’extérieur de l’INT 🙠… Mais ça y est, c’est a priori corrigé. Désolé pour cette annonce un peu rapide 😉

De retour des JRES 2005

Je suis de retour des Journées Réseaux (JRES) 2005, qui se déroulaient à Marseille, avec, je crois, plus de 1200 participants, où nous avons présenté notre papier sur l’intégration d’applications libres pour la réalisation de la plate-forme ProGET.

J’ai apprécié les présentations qui étaient d’une très grande qualité, en général, et démontrent une très grande professionalisation dans les équipes en charge des systèmes d’information et des infrastructures associées dans le monde de l’enseignement supérieur et de la recherche.

Le logiciel libre tenait une grande place dans les exemples de solutions présentés. Pour de nombreux acteurs, même si les utilisateurs finaux restent en deux environnements, Windows et GNU/Linux, l’infrastructure, elle est massivement équipée de solutions libres.

N’ayant pas suivi les développements récents des SI universitaires, j’ai été assez impressioné par la capacité apparente de mutualisation réelle des développements réalisés par les équipes de différentes universités. Au-delà des inévitables problèmes d’organisation ou de personnes, qui semblent avoir été résolus, on dirait bien qu’un certain nombre de projets libres, par exemple dans le domaine des Environnement Numériques de Travail (ENT) sont bien partis pour un développement collaboratif entre plusieurs équipes, et une diffusion dans de nombreux sites. Cela devrait permettre d’assurer une meilleure qualité des développements, une certaine généricité, et à moyen terme, une maintenance plus facile. Reste que ces projets seront plus ou moins franco-français pour la plupart… il y a quelques progrès à faire pour arriver à la dimension de vrais projets libres “classiques”. Mais je suis quand même assez bluffé, je dois le dire ;).

Vous pouvez consulter les actes des JRES avec l’ensemble des papiers et des présentations (Update : des vidéosde nombreuses conférences sont maintenant en ligne).

Gild : very intersting tool for introduction to software development with Eclipse

I’ve discovered a very interesting project, Gild, which was developped by collegues in the Victoria University, for the means of introducing students to both software development and the eclipse tool.

From the abstract of the paper presented at OSS2005 by Daniel German (actually I must confess I’ve had no time yet to read the full paper, and only attended the presentation and talked to Daniel…) :

This paper discusses Gild: An open source, Eclipse-based IDE for teaching and learning programming. Gild was designed to simplify and add pedagogical support to the Eclipse IDE to make it more appropriate for novice programmers and their instructors. Its development has greatly benefited from the ability to study, reuse, and modify existing Eclipse code. The core members of the Gild team are primarily researchers, making the maintenance of a growing code base difficult. It is challenging to create a community of developers because unlike most open source projects the developers (researchers) of Gild are not the main users (novice programmers) of Gild. To overcome this problem, we discuss techniques for making Gild more attractive to skilled developers (professors and graduate students). These techniques include improving instructor support in Gild and developing a grading perspective. We hope that these additions will attract able contributors and make Gild a self-sustaining community.

I find it very interesting as far we are concerned at GET, as we are also trying to teach computer science to our students using the same kind of approach : using libre software tools, and introducing them to the best practices of libre software projects.

We have actually adopted a similar strategy in developping a “scaled-down” sourceforge-like platform in the PicoLibre project, in order for students to discover these kinds of tools without spending the whole projects learning the intricacies of the platform. In this respect I think both tools, PicoLibre and Gild, are really very complementary : one concerning the environment of the project more on the server side, and the other focusing on the IDE on the client side. Our experienced is described in the paper PicoLibre: a free collaborative platform to improve students’ skills in software engineering by Éric Cousin, Gérald Ouvradou, Pascal Pucci, Samuel Tardieu that we had presented at the Second international IEEE conference “Systems, Man and Cybernetics of the twenty-first century”.

Not so surprisingly, both projects seem to suffer the same difficulties in maintaining a live community of contributors around the tools.

Maybe we’ll try Gild and adopt it… and who knows, maybe students of Victoria University will also work on PicoLibre in the future ?

Edukalibre and ProGET… two complementary systems

I went to the OSS2005 conference to present a paper in which we describe the work that we have done on the internal ProGET application (as part of the PFTCR project), a collaborative work environment for researchers at GET, based on libre software. In this paper we also outlined a strategy for contribution when using libre software, but that’s not the point here (no extensive description of ProGET is already really available, but the slides for our paper describe the main items : features and architecture).

At the same conference, a paper about the Edukalibre project attracted my interest, since the tools developped in this project are quite close to what we have or would need in the ProGET platform.

In ProGET we address users being researchers in a higher education school who want to work in a collaborative way with help of a web platform. In Edukalibre, the goal is to allow the development of educational resources in a similar environment. Both project address more or less the same users, then. And both projects are done by teams having a strong competence in libre software (and a commitment to this paradigm).

The features are very similar too : in Edukalibre, the documents repositories are using subversion for revision management and WebDAV (and other protocols/tools) for access to the documents. Whereas ProGET doesn’t yet use version management, we also provide a DAV interface, and subversion indeed was one option we were considering for doing version management in future evolutions. In ProGET we support some levels of confidentiality with respect to who gains access to which parts of the repositories, whereas in Edukalibre, everything is public. And we implement this through LDAP-based auth, when Edukalibre considers using this too for the next evolutions of their tools. Both projects are integrated with other groupware tools : Moodle components for Edukalibre, and PhpGroupware components of ProGET. When in ProGET we don’t recommend any document format, and include a wiki engine, Edukalibre supports mainly OpenOffice and DocBook documents, and enforces their use, and includes automatic conversion tools : both approaches to documents are really complementary.

Whereas what was done in ProGET is mainly internal, and thus not really a libre software project, we would very much like to be able to publish what we have done as libre software as a new version of the old PicoLibre platform.

As a conclusion, both systems are very similar and would probably benefit from collaboration, or at least offer to tools which can inspire each-other. Next step for me will be to test edukalibre 😉