Our thoughts are that if the device is inexpensive enough the trust and security issues go away because you can just send them home with the kids. Learning is much more effective that way in any case. That's why we designed the Afripad to be so cost effective. As for the hybrid model of live interaction + interactive content ... we are working hard at facilitating that in a very powerful way. And yes we are looking to use Android. I look forward to sending you what we have to solicit feedback when we learn the way forward. We look forward to you trying out both the Afripad and various tablets with the education management software we are developing. It will be in English but the content can certainly be translated.
By the way ... I really like your answer about live interaction being much more effective than asynchronous interaction for learners who aren't in the category in which their learning can be self-directed. But is live interaction the only type of interaction that can be effective? What about interactive learning content? In an environment of sometimes ineffective connectivity some kind of hybrid model might work well. There might be some activities that DEFINITELY need live interaction, but taking a hypothetical example, 1 hour of live interaction per day might be delayed by one or two hours when connections go down and still leave enough time for it to happen before the day ends. On the other hand a full day of live interaction might statistically never be completed unless the lecturer was available every day and could just pickup the next day where they left off. This is a hypothetical example that is not representative of what you are suggesting. I just outlined it to ask whether some hybrid live interaction/interactive content model might work well in your solution.