A Lesson Generation Framework to Enable Massively Distributed Content Development

Developing instructional materials to be used by e-learning systems is time consuming task. Systems -- especially those which do personalized lesson generation -- require the author to precisely follow a predefined pedagogic model while developing content. Careful coordination is required between contributors to ensure consistency and to prevent duplication of efforts. This slows down content development and also reduces the scope for reuse of existing digital content like video lectures. In this paper we propose a framework for a lesson generation system which supports massively distributed content development, by requiring only a minimum amount of prior planning and coordination between contributors. He/she only has to fill a simple metadata schema -- consisting of 3 sets of strings -- while adding the content to the repository. We show that, with a sufficiently large collection of such content in the repository, the system is able to perform the fundamental tasks of search, redundancy-elimination and personalized sequencing of content.