Managing electronic content for adaptation to the reader’s profile: project MultiAbile for the inclusion of impaired e-Learners

Project MultiAbile (http://www.multiabile.it) is an ongoing research and implementation effort to create a distance learning environment dedicated to the general public and to impaired people, offering a range of different channels and modalities to access the content and to study it. Some of the different modalities to access the content are: improved accessibility for screen readers; dynamic transcoding to synthesized speech over the telephone, with a dialogue-based navigation within the content; sound/tactile description of images (e.g. cartographic data) using force-feedback mice or pads. The content is also offered in two different textual versions, an original version and a re-processed version, in order to obtain high readability and understandability. The edited version is created in a way that only words from a basic dictionary are used (the Italian “Vocabolario di Base”), and the text is processed in order to obtain a high GULPEase index (an index of complexity in the structure of text). Users can describe their “profile” – whether they use a screen reader, or prefer to have a dialogue-based navigation, or if they prefer a simplified version of the text before confronting themselves with the original version. The user profile is used to dynamically adapt the content to the user, facilitating user access to it. In order to encompass the variety of channels and modality with which the user can “read” content in the platform, we devised a metadata schema with which organize the elements of the provided content, and a creation and editing workflow to process the content before publication in the learning environment. The resulting content is also structured in a way to be compliant to SCORM standards, so it is viable to process for multichannel and multimodal access also pre-existing SCORM contents. In the paper we describe the general data structure devised within the project to accommodate the different access modalities to the content, and to allow the user to access and read it dynamically according to his/her profile. We explain how this structure impacts on the workflow of creating and editing of the content, or of repurposing and adapting pre-existing content in SCORM-compliant format. As a conclusion, we advocate that this workflow and content structure can pave the way to the offering of the same content to different channels and with different modalities (including emerging channels, such as Digital TV, and specific modalities, like for example automatic rendition to sign language), limiting the effort of repurposing the content by means of automatic transcoding and transformation algorithms. 1 Foreword and Motivations: Electronic Content Accessibility The need to provide better accessibility to networked information is widely recognized and expressed in a number of initiatives and government regulations (W3C WAI, Section 508, etc.). However, most part of these guidelines focus on making more accessible the “navigation” or “publishing” environment (in the sense of the digital system through which the user can obtain the desired information, for example a web site through which the information is offered) more than facilitating the use and the comprehension of the electronic content within. Accessibility guidelines, in their current implementation, are mostly used as prosthetics, alongside other assistive technology devices (screen reading software, Braille displays, etc.), with web sites or publishing environments featuring an “accessible” portion alongside the “normal” part of the environment. While accessibility of the navigation/publishing environment is a necessary foundation, we advocate that much more attention should be devoted to the creation, aggregation and processing of electronic content, more than to the environment which is used to publish it to users. Improved strategies at content level allow to empower the user 186 Timothy Barbieri, Licia Sbattella, Antonio Bianchi, Ferdinando Carella, Marco Ferra Proceedings ELPUB2005 Conference on Electronic Publishing – Kath. Univ. Leuven – June 2005 offering him/her several alternatives to navigate and peruse the content itself, turning the way accessibility is currently implemented in electronic systems from a “prosthetics” notion to a truly multichannel and multimodal paradigm [2, 3, 12]. Focusing on how electronic content is structured, more than on the web site in which it “lives”, allows to dynamically repurpose it with transcoding and transformation procedures over different communication channels and interaction modalities (e.g. display on a web page with AAA-conformance, read aloud with TTS over the phone [4], convert into Braille or in an accessible/readable PDF eBook, provide information via a Digital TV screen). This way, once the user has selected the content he/she wishes to access, he/she can select the preferred channel and plurality of modalities. In a scenario of content adaptability over the channel chosen by the user, it is crucial that the user can express, in a formalized way, the features which are required in accessing the electronic content. This information, collected and organized in a general model, the user profile, is the base to describe the interaction modalities between an electronic publishing system providing contents and its user, impaired or not. In this paper, we present this approach, that we followed during the development of project “MultiAbile” (http://www.multiabile.it). We created a “publishing multimodal platform”, which can deliver electronic contents in different forms, based on a simple user profile described by the user himself. The content was described using a simple DTD devised with these objectives in mind [12], and aggregated following SCORM standards to provide the functionalities of a multichannel, multimodal e-Learning platform which could offer improved accessibility to the teaching material for the impaired learner. Fig. 1 – Focusing on electronic contents accessibility (right) over accessibility of the “publishing container” (left), it is possible to provide the impaired user with an adaptive, multimodal electronic publishing environment that better adapts to his/her needs, turning enabling technologies which are commonly used as prosthetics into real information channels. MultiAbile can be considered a “multimodal electronic publishing platform”, in the sense that it is designed to present electronic contents to its users using different channels and in different forms, according to the defined user profile. These forms are called “modalities” (and the platform “multimodal”), because content can be in principle accessed using more than one of its forms of representation at once [6]. Currently the channels supported by the implemented prototype are the web and the telephone, with possibility to transcode dynamically (with an XSLT-based approach) electronic contents to AAA-compliant HTML readable with Internet Browser, a Browser Screen Reader, or a Braille Display; VoiceXML with TTS/ASR navigation system accessible over the phone; Macromedia Flash cards with pictures, animations and colorful layouts. Over these channels, the user profile determines a number of modalities with which content is aptly optimized to better suit the user’s needs [10]. USER Electronic Content USER PROFILE C H