Creating Adaptive Textbooks with AHA! (An Interactive RoundTable)

AHA! stands for the Adaptive Hypermedia Architecture, an adaptive authoring and delivery platform developed as part of the Adaptive Hypermedia for All (or AHA!) project. (See http://aha.win.tue.nl/.) Adaptation in on-line textbooks makes it possible for learners to study the textbook in different ways without encountering difficulties. Link hiding and annotation guide users towards subjects they are “ready” to study. Adaptive sequencing or sorting helps them decide on a reading order for pages that teach them about a given concept. Conditional inclusion of fragments and stretchtext make it possible to provide additional or prerequisite explanations when needed or desired. In this roundtable (paper) we discuss the authoring process for adaptive textbooks created with AHA!. We cover the different authoring tools for concepts, concept relationships (like prerequisites) and adaptive object inclusion (for additional explanations). We also cover the process of completing an application by creating content pages in standard XHTML (or in XHTML+SMIL or SMIL 2.0) and by setting up a server. AHA! applications can also be created using other authoring tools, including Interbook (Brusilovsky et al, 1998) and MOT (Cristea et al, 2003), but this is described elsewhere. Introduction and Background Adaptive systems have started to emerge in the early 1990’s, with a strong emphasis on adaptation in educational applications. The overview paper by Brusilovsky (Brusilovsky, 1996) marked a turning point from mainly platform-dependent systems to Web-based environments. The updated overview (Brusilovsky, 2001) reflects this move towards Web-based systems. The work on the AHA! system (the Adaptive Hypermedia Architecture) started in 1996, and is still ongoing as part of the Open Source Adaptive Hypermedia for All project, partially funded by the NLnet Foundation. AHA! is attracting interest from many institutes and development effort from some, to add new features and tools , see e.g. (Cini et al, 2002), (Romero et al, 2002, 2003). The main goal of AHA! is to provide a general-purpose adaptive hypermedia platform, consisting of an authoring and a delivery part. In this paper we will concentrate on the authoring part of AHA! (but provide basic information on setting up the server part as well). AHA! concentrates on authoring at the conceptual level. Of course an author has to create (or retrieve and provide) pages or other content material, but AHA! first of all links pages or resources to concepts, and the authoring is then mostly done on concepts and concept relationships. Adaptation is specified at this level, and automatically translated (by the authoring tools) to lower level adaptation of links and content. In Brusilovsky’s terms (Brusilovsky 1996 and 2001) the AHA! delivery system offers a rich (and recursive) form of conditional inclusion of fragments and link hiding or link annotation . Contrary to most other adaptive hypermedia (learning) systems AHA! enforces very little in the adaptation or the presentation. We previously illustrated the presentation flexibility by creating a layout that closely resembles Interbook (De Bra et al, 2003b). The layout model allows detailed control over the presentation format. Concept relationship templates allow for the definition of adaptation to arbitrarily many different types of relationships, not just to prerequisites. Conditional object inclusion (De Bra et al, 2003a ) is a powerful means to add or select details or explanations (without redundancy) or to select between media types. Presentation stability enables adaptation without confusing the learner by presentations or information that changes on her when revisiting pages. This paper is organized as follows: we first show the overall architecture and typical aspects of authoring the content of an adaptive textbook for AHA!. We then introduce the authoring tools for creating the conceptual structure of an application. After this main topic we briefly describe some special tools AHA! offers for educational applications, including progress reports, forms for end-user manipulation of the user model, and multiple-choice tests. We conclude with a short description of advanced features of AHA! (mentioned above).