It would make sense if you could get electronic textbooks which are already in the question/answer form required by spaced repetition software, ready to "load and learn". Rather than reading a traditional textbook, you would take an equivalent question/answer set and learn it by spaced repetition. Do such things exist?
Following on from that, maybe such question answer sets could be automatically generated from the semantic web? A simple example would be a question/answer set about capital cities, but could one pick an arbitrary subject and manufacture a list of questions about it?