This extended abstract is a contribution to the Easy-to-Read on the Web Symposium. The contents of this paper were not developed by the W3C Web Accessibility Initiative (WAI) and do not necessarily represent the consensus view of its membership. For more information about RDWG symposia and publications, see the RDWG Frequently Asked Questions (FAQ).
The usage of electronic information and communication services is of great relevance in different parts of people's daily life, including private and work activities. People who are not able to use modern information technology - e.g. due to a disability - are therefore threatened to be excluded from modern information society. Especially people with intellectual disabilities regularly face problems preventing them from accessing information. These problems include hard or unfamiliar words, long sentences or complex phraseology. In addition an illogical or irreproducible information structure can prevent people of this target group from accessing information too.
This submission to the RDWG Online Symposium describes an ongoing PhD project that implements an approach using concepts typical for "Web 2.0" applications. The basic idea is to build up a community that composes explanations or easier to understand alternatives for hard to understand text elements. The community may create explanations using different formats including text, audio, video and images. These explanations are stored by a server forming a "glossary" service. An extension in the users' browsers queries this glossary and enhances the original web contents with these explanations where appropriate.
The research questions associated to this innovative approach for improving web accessibility for people with intellectual and cognitive disabilities are wide spread:
The described approach to adopt web content with additional data provided by the open community has been successfully applied in different projects e.g. Social Accessibility Project (Takagi et al.2008), Webvisum, Amara etc.. All of these projects have in common that their service focusses on people with sensual disabilities (blind and visual impaired, deaf or hearing impaired). This PhD project transfers this idea to a new application context respectively the new target group of people with intellectual and cognitive disabilities.
At this stage of development a first prototype of the service is implemented and publicly available at http://www.knoffit.de. It consists of a server and a client application (extension for Firefox browser). The next step in the project is to evaluate the prototype and to improve the service together with potential future users to achieve a User Centered Design.
To address the specific needs of people with intellectual and cognitive disabilities working with computer software, the decision was made to include the potential future users into the software development process. Because it would likely be too complex to let this group develop the intended service from scratch, a first prototype of the software was constructed without involving the target group. Instead the classic "Waterfall" software engineering model was used at this stage (Royce 1970).
To involve the future users in the further software development process, a qualitative survey has been developed. It asks the testers to accomplish certain tasks with the software, e.g. to install the glossary client in Firefox. By "thinking aloud" during working on a task the monitoring survey staff receives direct feedback and is able to notice possibly occurring usage problems (Someren 1994).
To solve recognized problems and to evaluate implemented solutions the software engineering model will be changed to an "evolutionary software process model" at this stage. This allows high frequent transitions between development and evaluation phases and hence, a continuously evaluation of the evolving final software product (Balzert 2008).
The final product will be evaluated to identify the usefulness of the proposed web service. At this stage of the project the method for this evaluation is not finally decided and therefore part of the ongoing research.
The main challenge of the project will be to develop and provide software that is usable and useful for the intended target group of people with intellectual and cognitive disabilities. The research approach plans to encounter this challenge by involving the future users into the software development process. Though Schäfers found out that qualitative work with the intended target group is generally possible, there are so far no research results available that cover qualitative work with the target group in the specific and complex application of software development (Schäfers 2008). In addition, webAIM acknowledges that there has been little research in the field of information technology and people with intellectual and cognitive disabilities at all (webAIM 2011).
The main outcome of this PhD project will be an internet based glossary service developed with and evaluated by the project's target group, people with intellectual and cognitive disabilities. It is to expect that knowledge gained during this project will be useful in other settings, too. The experiences made during the software development process with involved end users will be provided as guidelines, intended to help other software and web developers to produce better understandable and usable software or web sites.
The glossary supports production of explanations in multimedia formats (video, audio, images, hypertext). It will be interesting to analyze the effect of different media types on the understandability of complex words for people with intellectual and cognitive disabilities.
The current version of the software exists only as a first prototype and therefore it needs further development. The testing and improvement will be accomplished in close collaboration with the future user group. According to the project schedule the final version of the software will be available early 2013.
After finishing the continuously further improvement of the software a final evaluation is planned to prove its overall usefulness and its effect on understandability of web content. Especially the latter motivation appears to be challenging as it is complicated to "measure" someone's understanding of a text or the understandability of a text. Though there are readability algorithms and indices available (e.g. Flesch-Kincaid, Gunning-Fogg etc.) it is hard to decide whether a text is understandable or not. The concrete development of this final evaluation is part of the further research work.
I would like to thank the following institutions for contributing to the project: