Feasibility Analysis for SemiAutomatic Conversion of Text to Improve Readability

WCAG 2.0 (W3C Recommendation 11 Dec. 2008) recommendation includes a specific guideline regarding the readability of text contents. This paper presents an approach to adapt text to improve readability. We use existing guidelines for the development of easy to read materials together with natural language processing (NLP) methods to adapt text for people with learning or comprehension difficulties. The simplifications introduced address two basic problems: complex syntactic structure, and use of difficult lexical terms. Syntax structure is handled by obtaining syntactic parse trees of the sentences and applying a set of conversion rules to simplify them. Lexical complexity is addressed by replacing difficult words with simpler synonyms obtained from an electronic lexical database. We rely on readability metrics provided by existing online programs to identify problems and validate tentative solutions. After processing, the texts have been converted to accommodate the characteristics of people with special comprehension needs.