We propose a method of extracting and presenting appropriate sentences from the contents of existing IT textbooks for novices of IT technology. In our method, when a learner inputs a keyword, the system picks up appropriate sentences from all over the sentences of the textbook according to the following rules and displays them.: 1) The system determines whether or not a sentence should be displayed depending on the knowledge of the learner dynamically. 2) The system does not display the explanation which might be difficult for a learner to understand. 3) The contents to be displayed are classified into three types, basic explanation, detailed explanation, and applications. 4) The system explains necessary premise knowledge to understand the keyword. We discuss the possibility of the method. From an experimental result, it is expected that the method is successfully applied to the dictionary type e-learning system.
