OLPC : how about security ?
During the last week of February, I spent three days in Brussels, for the FOSDEM event. This is the biggest european meeting about free and open source software and it was the second time I went there (the first one was in 2005). The first track I had the chance to see, and not the least, was about the project One Laptop Per Child (known as OLPC). A project which has the ambition to propose a laptop for each child in the development world. The cost, 100$ per computer, would be paid by countries, societies and international organizations. A huge project in fact, because the product is not only seen as a way to make profit (for the campanies involved), but has for aim to provide a tool "desgned for learning learning". By this way, it might be able for them to get the knowledge to develop themself.
This project, as it seems firstly obvious, poses some ethicals and technicals questions. But, the more we dig these questions, the more problems we find. A lot have been, of course, strongly considered (the power supply, the screen, the networking capacity, and so on), but some aspects stay confusing. The aspect I want to focus on is merely how can millions of identical laptop be secured enough to avoid to be infected by viruses ? Just think that how it would be easy for a hacker if all the computers would have exactly the same architecture, the same softwares and, as a consequence, the same security holes. Here, the project OLPC offers the opportunity to spread viruses with a bewildering ease, as soon as a remote hole exists... which is not really absurd...
In conclusion, I see two observations. First, heterogeneity is good to fight against propagation of viruses. And finally, the relative difficulty for OLPC to be connected to internet, due to the bad network coverage of the development world countries, is maybe its better way to bother the propagation of viruses.