FREE SOFTWARE IN EDUCATION, SUCCESSES AND FAILURES Free software is slowly spreading in the education sector. Compare to a couple of years ago there are now a lot of deployment examples, we will look closely at some of them and examine their differences. From these examples we will determine a couple of key elements which could bring success or failure to such deployments. Finally we will speculate on possible evolution of free software in education. The GNU project was founded in 1984 by Richard Stallman while he was working at the MIT AI Lab. At this time, Richard started to face problem with closed source software their labs were provided with. He needed access to the source code of these programs to fix them and eventually he wanted to be able to redistribute his fixes. Until this time software were generally distributed in source code form so bugs can be quickly fixed. But time was changing and companies were starting to make business out of software, therefore distribution in the closed source form start to become more common. As a consequence, this form of distribution prevented the software engineers to fix themselves the bugs in these software. At this time, Richard Stallman decided to start the GNU project to produce an entirely free operating system. He also decided to frame the project in a legal context, he created the Gnu Public License - GPL - to legally secure the sharing spirit of the GNU project. A few years later the Free Software Foundation - FSF- was established. The GPL uses the copyright system to ensure free collaboration between developers. Richard was soon joined by several software engineers sharing the same vision. Gradually, with each part being written (GCC compiler, GDB debugger, Emacs editor, Bash shell, GLib standard C library), the GNU operating system started to become functional. As the GNU project started as an engineers project, its first users were software engineers, which in turn were also contributors to the project, this helps the project to grow quickly. In 1991, a young software engineer, Linus Torvalds started to work on the most missing part, the kernel. This part was named Linux, and this is why the Free Software Foundation prefers to call the GNU operating system GNU-Linux. Today, GNU-Linux is not anymore a software engineer operating system, it is now a good stable operating system grandma and grandpa can use for daily usage as browsing the Internet and sending email. This brings new users which, compare to the previous ones, will contribute little or nothing to the GNU project. As grandma and grandpa can use the GNU-Linux operating system, the children do so as well, therefore this clearly poses the challenge of using GNU-Linux in the educative sector. GNU-LINUX IN THE SERVER SIDE When Richard Stallman started the GNU project he wanted it to be a UNIX clone. UNIX was designed for network and time sharing. Gnu-Linux shares the same quality however the GNU project attracted a lot of developers to participate to cross-UNIX projects like Samba. Samba is a suite of software for UNIX alike system to emulate the functionalities of a Microsoft Windows server, but without the cost. In the late of '90s it became very obvious to several network administrators in schools that GNU-Linux with Samba could replace a Windows server without the license seat cost. As these administrators package the component they also introduce new ones as proxy system, user administration tools, print sharing system, mail services, intranet services,... Several of these projects are developed and supported by governmental agencies. In France we have several projects like that: - Pingoo [1] is supported since 1997 by the local government of Haute-Savoie[2]. It is deployed in hundred of primary and junior high schools, town halls, tourism offices. The Pingoo servers are administrated and updated remotely from the Haute-Savoie computer resource center[3] - SLIS[4] is developed by the Grenoble academia agency[5] and deployed in hundreds (more than ~700) schools in this area. SLIS is only a communication server, a router to the Internet with filtering capabilities. - EOLE[6] is supported by French education office. - SambaEdu[7] is developed by the CRDP of Basse Normandie. It is part of the SCEREN/CNDP network[8] a national agency related to pedagogical resources and ICT in education. These projects share a lot of common characteristics: - Administration of the server is expected to be centralized. - No or little local administration is expected on the server, local administrator does not (and can not) administrate this server. - These solutions share the same categories of functionalities: at the minimum an Internet-LAN router with filtering capabilities but it is more common to have this server to behave as drop-in replacement to a Windows server with file and printing sharing services. - The workstations are expected to be Windows ones. These deployments are successful mainly because: 1. the distributions are designed for a limited scope 2. the deployments are carefully planed, with the assorted needed human resources 3. no specific local school knowledge is required, the installed server are like black box which can not be accessed 4. the proposed functionalities are limited 5. possibilities of extension are low, in most case only software update 6. the workstations are Windows ones From the user point of view, these projects does not disturb usabilty behavior, so it is mostly invisible to the user: he will keep the same habits. GNU-LINUX IN THE WORKSTATION SIDE GNU-Linux in the workstation side is a much bigger challenge as the end user will directly interact with the GNU operating system. Nevertheless some pioneers started to promote this approach in the late '90s. 1. ABULEDU[9] It started in France at Bordeaux under the inspiration of a primary school headmaster. He described to a local free software group the features he wanted for its ICT infrastructure. The solution proposes a server with the same functionalities as before plus a thin client system to operate GNU-Linux workstation. The interesting aspect is that a thin client can be a recycled computer. With a thin client, the application run remotely on the server, the thin client is only responsible to display windows and to send mouse and keyboard events to the server. This solution helps to equip school at very low cost and facilitate the workstations administration. This project started as a full private one, later it got support from local government to pay partially a software engineer to conduct the job. An interesting note: later the project was updated and improved under a corporate form[10] by the initial AbulEdu software engineer, however the project itself is not reverse back anymore to the community as it is only shipped with new purchased server. As a consequence, the server can be sold with professional support. but the initial project, without sustained development effort, slowly became obsolete, although the estimate user base is about 30000 (including the corporate version). The AbulEdu started as a community effort of the civil society, latter it emerged as a corporate effort, which is helpful to bring long term vision to the project, sadly it is not anymore in the community. The public sector played little in this project. Schools willing to use AbulEdu community are generally on their own and they get in most case no support from the public institution to install and to train the teachers. 2. DEBIAN-EDU This project emerged in Norway, then it integrated the Debian project as the education sub project. While newer than AbulEdu it shares the same functionalities, it is gradually becoming an international effort. The project has support from Norwich NGO to pay software engineers and developers gathering. 3. LINEX[12] and GUADALINEX[13] These two projects from Spain are initiated and supported by the local governments of Extremadura[14] and Andalucia[15]. Their goal is to bring GNU-Linux in the desktop in schools, town halls and a other public administrations. This is generally massive deployments. They designed their own distribution (both are in fact quite close) with a specific user interface. WHAT WE SHOULD REMEMBER There is a lot of other projects with goal to bring GNU-Linux in the desktop. These projects can be divided in two categories: ones coming from the civil society with no particular support from the public institutions and other supported by public institutions. The first one relies mostly on volunteer resources to install, to maintain, to train the users. Such situation jeopardize the stability of the installation. Eventually a school can buy professional support from local companies to do these jobs. The second one 'forces' uniformity in a large scale. The immediate benefice is to facilitate knowledge exchanges in the user base. Indeed as GNU-Linux is completely open there are hundreds of possible Desktop arrangements a user can face. Then Desktop user experience can be very different from one GNU-Linux to another one. Forcing to one Desktop environment facilitate deeply the training of the users and exchanges of know-how and tricks & tips. A massive deployment facilitate the planing to install, maintain and train the end users. It also helps to bootstrap a local economy on ICT services and training. It will be very interesting to follow the evolution of such projects and their economic consequences in the coming years. Nevertheless, uniformity is only on the surface. Indeed one of the weakest point of the GNU-Linux desktop is the multitude of user interface paradigms. For example to save a file, depending on the application, a user may face different user interface to do that basic job. Try to do it with OpenOffice, Firefox, TheGimp, Scribus, Audacity, Kstars, TeXmacs, Dr. Geo; you will face 8 different dialogs to do it, with different layouts and behaviours. From a user point of view it is a very expensive experience with a lot of stress. This problem is endemic to the open nature of the GNU-Linux desktop and the users need to be prepared to face it. Deploying GNU-Linux Desktop in the school is very risky. In France it is very interesting to note there are several competing server side GNU-Linux projects supported by public institutions but none for the Desktop. Proposing GNU-Linux in the Desktop means a lot of change to the users. Deploying GNU-Linux Desktop means taking risks which the politics may not want to take, even if it means a better use of the public money. Such choices may also put the educative media industry, which is Windows oriented, in a weak position. When a school is installing GNU-Linux desktop without external support, they should have internal human ressources able to administrate the machines and to provide initial training to the users. Some school are able to pay for external services to install and to administrate GNU-Linux workstations. Without these kind of support it is very hazardous to deploy GNU-Linux Desktop. With these recent deployments as in Spain, it appears the time of the pioneers is over. Instead the GNU-Linux Desktop becomes institutionalized. Even further, Andalucia and Mike Shuttleworth's Canonical Ltd will work together for the next versions of Guadalinex[16]. This clearly shows that GNU-Linux starts to penetrate the workstation segment. The next challenge is the ability to provide free quality educational software. Designing quality software takes years (for Dr.Geo it took several years during my free time), the software been free or proprietary does not change the complexity of the developing task. Compare to other software categories, few developers are interested by educational software, therefore incentive solutions need to be find to bootstrap and sustain such efforts. The model to achieve this has still to be invented but a general idea would be to get supports from institutions deploying free software for education. One may say that with massive deployment in education of the GNU-Linux desktop, the community may gain more developers, I am tempted to respond yes and no. 'Yes', the number of users increase, 'no' these new users are not software developers. GNU-Linux started as a software engineer project, they were the first users and at the same time developers and they contribute a lot to improve it. However this is not the case with new users being educators, very very few of them may develop free educational software, mainly because they do know how to do it and they do not have the time to learn it. FREE SOFTWARE UNDER NON-FREE OPERATING SYSTEM It is not as spectacular as deploying GNU-Linux desktop, but it is a first very interesting step. In area where you know you can not deploy GNU-Linux desktop, it is still possible to deploy free software for Windows. In the local area of Landes[17], a special operation is taking place since several years, junior high school students get for the school year a notebook they can bring home. While equipped with Windows, we also provide software like OpenOffice.org, The Gimp, Dia, Audacity, Squeak, Mozilla, etc. [1] http://www.pingoo.org [2] http://www.cg74.fr [3] http://www.cri74.org [4] http://www.ac-grenoble.fr/carmi-internet/slis/ [5] http://www.ac-grenoble.fr [6] http://eole.orion.education.fr/ [7] http://www.crdp.ac-caen.fr/se3/ [8] http://www.sceren.fr [9] http://www.abuledu.org [10] http://www.abuledu.com [11] http://www.skolelinux.org [12] http://www.linex.org/ [13] http://www.guadalinex.org [14] http://www.juntaex.es/ [15] http://www.andaluciajunta.es/ [16] http://www.guadalinex.org/modules/news/article.php?storyid=132 [17] http://www.cg40.fr