Proposal from OLPC Paraguay on how to manage Sugar or other educational software
The project to deliver One Laptop Per Child (OLPC) for educational purpose in developing countries is doing great in Paraguay. According to developer Bernie Innocenti, this success comes from a way to manage the development of the Sugar educational software that other countries (or any other similar projects, see for example the Teachermate or the italian JumpPC) could and should imitate.
Stop: Bernie, what do you like in OLPC Paraguay?
Bernie: The fact that, even though it's a little deployment it contributed to the Sugar educational software much more than other countries that have 100 times more resources. Since when we started we've moved from passive dependency from others to full participation to the international Sugar community, overcoming language and cultural barriers.
Stop: Is Paraguay the only country where this happened?
Stop: What's the secret of this success?
Bernie: When there are enthusiast managers and people with the right attitude it's easy to do well. This said, the success in both Paraguay and Nepal is a direct consequence of the fact that everybody, from end users to developers, works together in close contact to continuously improve the system.
Stop: What would you like to do next?
Bernie: Take all the work done by the OLPC and put the last version of Sugar (see here for details) on top of it.
Stop: Is that a complicated task?
Bernie: It isn't, from a technical point of view. The problem is that in order to do that we should slow down our work and get back in sync with the main project. OLPC prefers software released less frequently but with long term support. In this period they're deploying an upgrade based on Fedora 11 and Sugar 0.84, two products released more than one year ago that are near their end of life.
Stop: So it's a management problem?
Bernie: Right. The business mentality isn't compatible with a community-centered development model. This also makes it harder to be trusted by the bigger OLPC deployment, like Plan Ceibal, that distributed in Uruguay 400 thousand laptops to all the kids in primary school.
Stop: Why? How do they work?
Bernie: They work like businesses, trying above all to get support contracts. I heard, for example, that Plan Ceibal wants to exploit all the know-how that they put together with great effort to offer paid support to other OLPC deployments. Fact is that they didn't contributed almost anything in two years and are still stuck to Sugar 0.82 which is two years old. Goold luck to them! Working in isolation on an obsolete version will make it difficult to find people interested in paying for support.
Stop: Could you make some concrete example of the difference between your ways of working?
Bernie: The support for GSM modems developed by OLPC Paraguay is one of the most important additions to the last version of Sugar and, as it's the case with all our software, everybody can download it from the Internet under a Free license. The source code of most of the Sugar customizations made by OLPC Uruguay, instead, is not available online (most of those changes are aimed at controlling the kids or limiting how they can use the laptops, in contrast with the Free Software spirit and the OLPC Child Ownership principle, but this is another issue). Luckily, there are exceptions.
Stop: Like what?
Bernie: When we realized that we needed some backup and restore software for the Sugar Journal we asked for it to Plan Ceibal because they had already done it. Initially they just answered that they would need to contact their managers, so we decided, even if it would have been had an absurd waste of time, to rewrite the same software ourselves. Luckily, however, a few weeks ago they released their software online. I really appreciate this move a lot, since it shows they're willing to change.
Stop: So, what's your proposal to work more effectively on Sugar?
Bernie: Maybe it would be more effective to try to hire programmers who are already well known in the community, insist that all patches and improvements are always made available upstream (that is to the original programmers) and finance the most critical projects that are stuck for lack of resources.
Stop: Considering that Sugar can run almost on any computer, not just on the OLPC laptops, I hope that's just what will happen. Keep up with the good work, Bernie!