Real and exploitable knowledge, free software is a means for a scientific researcher to promote and disseminate his discoveries in society. This is the path advocated by Danielle Le Berry, teacher-researcher in computer science.
Firefox, OpenOffice, VLC, Ubuntu… Like these few famous examples, free software — software whose license allows everyone to use, modify and redistribute — often plays key roles in our digital ecosystems. We find it on our servers, in our TVs, on our phones, on our computers… Good news: it should last a long time. Actually thank you Digital Republic Law for 2016 and The second national plan for open scienceSoftware generated by publicly funded research should by default be distributed under a free license.
It must be said that the academic world and the world of free software are closely linked: the GNU operating system project was created by Richard Stallman in 1983, when he was a researcher at the Massachusetts Institute of Technology, Linus Torvald developed the Linux kernel in 1991 while he was a student at the University of Helsinki… This remains relevant because many free programs are developed or improved in our laboratories. On February 5, a 1st Prize “Open Science of Free Research Software”, Organized by the Open Science Commission of the Ministry of Higher Education, Research and Innovation, it proved this by rewarding ten programs from the French research world, out of 129 quality applications. Many of these applications were for software developed over 30 years ago!
Software is, in a way, “actionable knowledge”. By authorizing its distribution under a free license, we remove any obstacles to its use and encourage the adoption of this knowledge in the academic world and in society in general. Moreover, in the field of dematerialized software, such distribution can be done at a very low unit cost. Thus, the principle of openness by default is an opportunity to further enhance the activities of our laboratories. This promotion can take several forms: adoption in the community, in contexts different from those for which the program was designed, or participation in maintenance efforts.
In the world of research, software is created to meet a specific need, or as proof of concept for new scientific developments. However, in the world of free software, users can also be contributors: report bugs, suggest improvements, modify code directly to correct or improve the program. Thus, increasing the visibility of free research software means increasing the number of potential contributors, improving their maintenance and increasing their functional and technical scope. It is therefore a source of innovation, allowing software to be used in a very different context from that for which it was initially created.
Free software, a community story
All developers, users and contributors constitute the free software community. When the software community becomes large, a way must be found to organize it and develop its economic model. If a small company can be moved with limited means, then a large company requires structuring and allocation of human resources. It is obviously not enough to put the code of the program on the Internet under a free license for the automatic creation of a community. The program may take time to find its audience. The decision to distribute the software under a free license comes down to taking into account that this software, which has been useful to us, can also be useful to others, and therefore has a means of interacting with its users. From the user’s point of view, the adoption of software remains a matter of trust: the user considers that the software does what it should do, and that it has been developed according to the rules of the art ; The user also hopes to find help in case of problems, especially when he encounters an error. In the case of free software, this trust is based in particular on the examination of the code itself or on interactions with other users or contributors.
Therefore, any development approach must consider interaction with users and contributors, openness of the development process, and transparency of program development and governance. The simplest framework is open development, using a fake (like Github or Gitlab for example). These counterfeits make it possible to access the code and manage comments and user contributions, making it easy to follow its evolution and participate in its development.
Linux exhibition at La Défense, Paris, May 2011.
Another solution consists in integrating an already organized free community (FSF, Apache, Eclipse, OW2). These communities, which benefit from recurrent funding, offer more services (mailing lists, forums, websites), within a more structured framework (governance, development practices, copyright management, etc.). There are two European free software communities that can host free software research: OW2 and [email protected] These communities also bring their expertise in the management of free software in organizations (see for example OSPO.zone for OW2/Eclipse).
Blogs for Respect
Promotion by free software must respect the habits and customs of the free software world. If several members of laboratories know them individually, they must be combined in a more structural way to reduce the effort necessary for this type of evaluation. In particular, at the laboratory level, it is necessary to simplify the procedure for declaring programs, developing dissemination strategies for public or institutional fraud, promoting this program using the channels specific to each community (language, domain ).
There is therefore a step to take between providing software under a free license in the framework of open science and actively developing it as free software. This fundamental step is the acceptance of interaction with the community, and requires the implementation of the means that allow it. This activity requires special skills that must be developed in the laboratories. ♦
The views, opinions and analyzes published in this section engage only their authors. It does not in any way constitute a position of the CNRS.