Intelligent Information and Communication Systems
English version deutsche Version|

The Logo of the BenToWeb project

The acronym BenToWeb stands for "Benchmarking Tools and Methods for the Web". It is the name of an EU funded research project in the Sixth Framework Programme's (FP6) Information Society Technology (IST) priority. All in all nine institutions, public and private, from all over Europe, take part in this project. Please click here if you would like to learn more about the BenToWeb project.

The aim of the project is to develop methods to automatically control criteria for accessible web pages. The Web Accessibility Initiative (WAI) of the World Wide Web Consortium (W3C) produces accessibility criteria for web pages. These criteria are laid down in the Web Content Accessibility Guidelines (WCAG). An overview of the WCAG can be found here.

The team at the IICS of the FernUniversität in Hagen work on the parts of the project that deal with natural language contents of web pages. Two questions are central to our work:

  1. Which are the criteria that must be met for a text to be easily understandable?
  2. Which of these criteria can be controlled automatically?

An overview to both questions is given in the deliverable D6.1. This publication and more publications concerning this project can be found at http://pi7.fernuni-hagen.de/publikationen/delite.html.

Screenshot

Currently we developed a prototype named DeLite. This system determines on the one side a total readability score which indicates the understandability of a text on a scale from 0 to 100. On the other side it shows on which levels (morphological, lexical, syntactic, semantic and discourse) text passages could be found which are difficult to understand (see screenshot).

If one of those levels is selected, using the mouse, the associated difficult text passage will be highlighted in the appropriate color of the level. In the example, shown in the figure above, the text is difficult to understand, because the German pronoun er can relate to either Dr. Peters or Herrn Müller. Except "Number of possible antecedents for a pronoun" we judge the readability of texts by several other indicators, e.g. sentence complexity, number of syntactic ambiguities or double negations.

Hermann Helbig | 04.03.2009
| Valid HTML 4.01 | Valid CSS 2.0 | Level Triple-A conformance icon, W3C-WAI Web Content Accessibility Guidelines 1.0