The sustainability of free/open source software

The paper applies three analytical frames to a better understanding of the itch to invent and innovate cooperatively, a still inadequately treated stylised fact, while drawing some lessons from an ongoing Free/Open Source software project on communication standards, software and services: Jabber , taken as an eloquent case and test bed for the proposed three - layered frame. The first frame derives form the territorial innovation systems literature: some features of the Internet economy, and particularly such standard - setting institutions as IETF working groups, provide a favourable climate to the governance of cooperative software projects. The second one is drawn from the economic theory of networks: the actual inducement s to cooperate can be explained by a class of models about the incentives and costs faced by an agent, rationally deciding whether to join a network and betting upon choosing a fitter one. The third one improves the latter, by introducing a simple evolutionary frame: the software project lifecycle. On the analytical level, a major finding is that economic models overestimate "cooperation failures": if developers were strictly "rational", they should cooperate at a much lower scale compared to observed patterns. This puzzle leads to the suggestion of re- introducing Smithian Moral Sentiment s into economic analysis. As another major point unveiled from the evidence of the case is the sensitive insuppressible key role of intrinsic motivations in this kind of innovative enterprises, linked strictly with the core nature of free /open source style of organizing. It stems that, in terms of institutional arrangement s, there's a wide spectrum of possibilities to experiment with, taking absolutely care not to destroy the vitality of the free ecology mining the critical drives of the innovator s. As far as policies are concerned, the paper aims to switch our attention to the long term sustainability of the novel software - services business models, and a "just" distribution of collective innovations net benefits.Le papier applique trois cadres analytiques pour essayer de mieux comprendre le phénomène de la création coopérative dans le cadre du logiciel libre tout en tirant quelques leçons d'un projet libre sur les standards de communications : Jabber. Le premier cadre fait référence à la littérature sur les systèmes d'innovation territorialisés : certaines caractéristiques d'Internet sont favorables à la gouvernance de projets logiciels coopératifs. Le second cadre fait référence à la théorie des réseaux : des modèles récents tendent à expliquer comment et pourquoi des acteurs choisissent de joindre un certain réseau. Le dernier cadre fait référence au cycle de vie du logiciel et permet de comprendre quelques faiblesses du modèle du libre. La conclusion présente quelques faits stylisés à évaluer dans de prochaines recherches empiriques