Well, indeed it's something to consider. As mentioned above, i've got it in my head that the "prototype business model", can we say, would be through a university, and would fit into a regular semester. That's the only drawback of adding extra complexity by virtue of being a web-app. That, and the fact that kids in Malawi, or Kenya, or many many many other places, don't have high speed internet, and, much of where i'm coming from has been influenced by my 3 yrs spent teaching in Kenya (
U.S. Peace Corps) -- in a school without windows or chalkboards, much less telephone, much less internet. But i could definately see those kids getting into stand alone education applications developed through a global coalition, based at schoolforge, running on OLPC machines, etc....
Here is a link to a graphic describing how i once sought to implement such a program at NMSU: http://www.asymptopia.org/images/pea.png