The Internet has become an important learning tool for both professionals and students. While search engines have improved the ability of learners to find educational materials, they are still lacking in many aspects. In many cases, search engines do not provide relevant content, and users have to undergo an exhaustive and iterative search process to find it. This paper presents a novel approach for automatically generating an e-textbook from a user-specified topic hierarchy, thus improving the learning experience for the user.

According to the authors, the generation of an e-textbook starts with a user-specified topic hierarchy. The process of building an e-textbook is then made up of four steps: dataset collection, which is used to generate different queries to search the Internet, and gather the results by using a search engine; mining, which is used to process the retrieved Web pages, and find suitable pages among them; expansion, which is used to expand the results list for nodes that need additional pages; and result presentation, which is used to display suitable Web pages.

The paper addresses an important limitation of search engines and e-learning, and therefore is a significant contribution to the field of e-learning. The paper uses existing techniques, and introduces new techniques, in its process steps, and clearly explains the process of building an e-textbook.

I agree with the authors that starting with a concept tree and building an e-textbook based on that tree is a good model, and works well with modern integration approaches. However, I believe that it is not easy to develop the concept tree on one’s own, even for professionals. The authors should have addressed how to aid users in developing the concept tree. Liu et al. [1] present an iterative approach, referred to in the related work section of this paper. This iterative approach, combined with the authors’ notion of explicit definition of the concept tree, might be a better solution.

Another important aspect of developing e-learning systems might be the use of user profiles. For example, if a user has subscriptions to different journals, then the e-learning system should factor these subscriptions into its search. Otherwise, the search results might not be comprehensive enough to address the needs of the learner.

In the result presentation step, the authors describe the results as presented on an interface similar to a traditional book. Although this could be a good starting point for the result presentation, many new Internet techniques could be employed to make e-learning a much better experience for the user than a traditional book.