Web Design for Dyslexics: Accessibility of Arabic Content Areej Al-Wabil, Panayiotis Zaphiris, Stephanie Wilson Centre for Human-Computer Interaction Design City University, London EC1V 0HB [email protected]
, [email protected]
, [email protected]
Abstract. This paper reports results of a workshop on the design of electronic content for users with Specific Learning Difficulties (SpLD), particularly Arabic dyslexics. First we shed some light on the nature of the Arabic language and discuss features that account for the unique needs of Arabic users with reading disorders. Then we present recommendations for accessible web design for Arabic content in light of existing guidelines on web design for dyslexic users.
1 Introduction The use of electronic materials is increasing and more educational content is being made available to users online. Web design guidelines promoting accessibility for users with Specific Learning Difficulties (SpLD), dyslexics in particular, have been developed (BDA, 2004; Evett and Brown, 2005). Until recently, our understanding of dyslexia was largely confined to research on the English language. For this reason, existing accessibility guidelines fail to address some issues related to non-Latin orthographies. For accessibility issues related to presentation and content, it is important to consider the linguistic properties of the content, especially for the case of users with SpLDs. Evidence in recent studies has shown that linguistic properties of different languages such as phonological, morphological, and orthographic characteristics influence the learning of reading and writing of that particular language (Goulandris, 2003). A few studies have begun to appear that look into characteristics of dyslexia in Arabic (Abu Rabia, 2004; Beland and Mimouni, 2001; Elbehari, 2004). Design guidelines for Arabic content need to consider features of the Arabic language that account for the unique needs of users with reading disorders. Arabic is widely spoken with an estimated 200 million native speakers. It is also the language of the Quran and is believed to be the second most widely used script in the world (Holes, 1995). The spread of Islam prompted the incorporation of other cultures outside the Arab world and resulted in what is referred to as the "Arab scripted world", which includes contemporary languages such as: Farsi, Jawi, Kurdish, Pashto, Sindhi, Urdu, Dari, and Uyghuri (Milo, 2002). Issues related to the design of Arabic electronic content were explored in a workshop investigating accessibility of Arabic content for users with SpLDs, particularly dyslexics. The objectives of the workshop were to identify the most commonly noted difficulties which dyslexics experience when reading and writing Arabic script, understand the needs of dyslexics for accessing Arabic electronic content, and present recommendations for accessible design specific to Arabic content.
1.1 Arabic orthography Arabic is an alphabetic language, written from right to left in cursive form, with 28 letters. Each letter has a different form depending on its position in the word (initial, medial, final). Arabic script includes long vowels, but no short vowels. Short vowels are indicated by diacritical marks, small diagonal strokes above or below characters. The Arabic orthography is either shallow (short vowels appear in the writing as diactritics) or deep (unvowelized, diacritics are omitted). Both orthographies are used for electronic content made available online. However, deep orthography dominates the majority of websites. While newspapers and textbooks utilize deep orthography aimed at skilled readers, shallow orthography is used in children's books and in formal texts such as the Quran (Abu-Rabia, 2004). Texts presented in the shallow orthography are easy to phonologically decode due to the consistent letter-to-sound correspondence. Very few websites provide content in this type of orthography for two salient reasons: First, a greater expansion of effort is required from the developer to include diacritics, especially for larger websites. Secondly, the linguistic skill required for the correct application of diacritics can be intimidating even to native speakers working as developers.
Arabic, like other Semitic languages is also bi-directional (BiDi for short). Most of the text is written from right-to-left, while some of the text (like numbers) is written from left to right. This requires special accommodation for editors and browsers and only recently have products been able to address this issue and facilitate bidirectional support. Arabic is also a highly homographic language where the same word can carry many different meanings (Abu-Rabia, 1998). 1.2 Dyslexia in Arabic readers Very few studies have investigated dyslexia in native readers of Arabic. Common problems of dyslexia that have been observed and reported in the literature of Arabic dyslexia include (Adapted from AlSarhan, 2001 and Al-Sartawi et al., 2001): Letter and word recognition for letters that are similar orthographically or phonetically, letter and word omissions and additions (including diacritics - short vowels), short term memory problems, letter and number reversals, and disturbances in visual and audio processing. Due to the idiosyncratic nature of dyslexia, these difficulties afflict individuals in varying forms and degrees of severity. Several studies by Abu-Rabia (2001, 2004) have investigated dyslexia in Arabic and showed similar characteristics to those in other languages. These studies have shown that Arabic dyslexics have poor phonological processing skills and working memory (cognitive processing) compared to normal readers. A significant difference is the visual processing required for diacritics in Arabic (Abu-Rabia et al., 2003). Elbehari argues that factors such as the cursive nature of the Arabic script, inflectional morphemes, different graphemes, and discrepancy between phonology and orthography in unvowelized scripts could possibly be contributing to the manifestation of dyslexia in monolingual Arabic speakers (Elbehari, 2004).
2 Methods A workshop on assistive technology for users with specific learning difficulties related to reading and writing of the Arabic language, particularly dyslexics, was held in Riyadh, Saudi Arabia on January 3, 2006. The objective was to review guidelines promoting accessibility of electronic content for dyslexic readers and explore issues related to Arabic content. A total of fifteen participants included specialists in specific learning difficulties and educational software development. Four participants were experts in SpLDs from the college of special education in King Saud University, four specialists in SpLDs represented the General Secretariat of Special Education, four special education teachers experienced in working with dyslexic students at Ministry of Education public and private schools, and three computer scientists with experience in developing Arabic software from Prince Sultan University and the Ministry of Education. Issues affecting the accessibility of Arabic electronic materials such as readability, user control over presentation, and information architecture were discussed. To address the objectives of this workshop, an Arabic translation of the checklist for web design for Dyslexic and Visually Impaired readers (DYVI) was reviewed (Evett and Brown, 2005). Each item/guideline was presented with an example of an Arabic website or electronic content. For each guideline, we discussed the underlying dyslexic difficulties and their design implications. Furthermore, we asked participants to rate the significance of each guideline with relation to Arabic electronic content on a scale of one to five, one being not significant and five being very significant. Other issues, specific to Arabic script and not addressed in the published guidelines, were also raised and discussed in our review.
Results and Discussion
Participants were in greatest agreement about the benefits of computer-based materials that could provide content that is configured to users' needs. The capability of providing alternative methods of content presentation such as audio and video was believed to be helpful for dyslexic students. Audio, in particular, was reported from participants' practical experience (due to the availability of audio recording of some text based materials in the curriculum) to be a helpful aid for students in understanding content. It has been noted that for many users with SpLDs, "the accuracy of the perception of the spoken word is greater than the perception of the written word" (Evans and Blenkhorn, 2005). It is noted that many educators appeared to be unaware of how simple changes to the interface could affect the ability to read, as has already been demonstrated in studies on users of the English language (Gregor et al., 2003;
Kurniawan and Conroy, 2006). In the workshop review of the DYVI checklist, most of the items were found to apply directly to Arabic content and were rated very significant with the exception of a few. For example: visual contrast, minimum font size of 12pt, and avoiding light text on dark background had the least average rating among participants. It was also noted that some items, in the DYVI checklist, do not apply to Arabic content such avoiding capitals mid-line and using full stops for abbreviations and acronyms. It was agreed that alignment of text should be from right-to-left due to the nature of the script. However, fully justified text was not seen as a problem because in Arabic, words can be stretched to fill the width of the line instead of introducing inconsistent spaces between words as in English. This type of justified text was reported by participants, from practical experience, as being quite helpful in letter recognition and associated diacritics, as well as providing a visual aid when reading from right to left and moving from one line to the next. As for font types, participants rated the clarity of scripts that have flowing cursive form with full and deep curves and straight and vertical uprights such as Naskh much higher than angular types such as in Koufi and Andalus types. The preferred text font types, rated by participants, were Arabic Transparent followed by Simplified Arabic Fixed. As for text size, the preferred size was in the 16-20pt range. This is not surprising as Arabic script is usually presented in larger sizes than English, both in print and electronic form. In terms of emphasizing text, results were in agreement with the recommendation of using bold rather than italic or underline. Many suggested the use of color or a combination of both bold and color as is used in many Arabic textbooks for early elementary levels. This is also in agreement with the fact that "Adding color can increase the likelihood that the information will enter the long term memory" (Rainger, 2003). The recommendation of using short sentences was given an average 4.7/5 significance rating by participants. It was agreed that shorter sentences are easier to read and comprehend. However, adhering to this guideline might present a challenge, as the nature of the Arabic language allows for longer sentences that span many lines, a distinct contrast to English. Spacing between letters and words is a problem in the presentation of Arabic texts. Six letters in the Arabic alphabet cannot be joined, and consequently create spaces between letters within words in Arabic. Readers must learn to distinguish word boundaries (space between words) from breaks in the cursive script within a word. This might add to the difficulties dyslexics face in the visual processing of text. It is recommended that spacing between words be increased. Perhaps the most interesting point which arose during the review of the guidelines was the significance of diacritization (vowelization) when promoting accessibility of Arabic content. Participants agreed that the use of diacritics could improve reading accuracy and comprehension, especially for beginning readers. In a study on the role of vowels in the reading of Semitic scripts, Abu Rabia (2001) states that "If texts are not vowelized, these short vowels must be deduced by the reader, which is a cognitively demanding task". In another study (Abu-Rabia, 1998), he reports: "Results indicated that vowels had a significant effect on reading accuracy of poor and skilled readers". This is not an issue in English, as vowels are present in all text. For example, consider the English words "heard", "herd", "hard", and "hired". A deep, Arabic orthography would reduce these words to "hrd" and then leave it to the individual reader to decipher the word and add its missing vowels depending on the context. Furthermore, the extensive use of deep orthography in Arabic electronic content adds another barrier to accessing its content to both visually impaired and dyslexic readers. The visually-impaired group is affected because screen readers are prone to mispronounce words that do not have diacritics, thus rendering content incomprehensible. The dyslexic group is also adversely affected by being forced to read without diacritics, because without them they must conjure a combination of contextualized cognition and visual memory, thus adding to their cognitive burden. While diacritics help to disambiguate words, they also have the negative effect of adding to the visual complexity of the script. Reading with diacritics requires high levels of visual discrimination and memory skills. This in turn reduces the reading comfort of both beginning and experienced readers. A suggestion was made to offer text in both formats in electronic form and give the user the option to view or hide the short vowels. Offering levels of vowelization, full or partial, would address the visual complexity of the script. Some assistive technologies have been developed to automatically add diacritics, such as Sakhr's Automatic Diacritizer. However, this solution is specific to a limited range of commercial products and not affordable to most users. The Unicode standard accommodates the representation of diacritics. It is recommended to consider automatic diacritization of text in applications so that users can control the appearance of diacritics in electronic Arabic content.
This paper presents recommendations for promoting accessibility of Arabic content derived from practitioners' experience. Suggestions were made to offer diacritization to facilitate comprehension and to consider simple and clear Arabic typography in presenting text to increase readability. These recommendations need to be further supported by usability research and tested experimentally at a later stage with dyslexic users of Arabic content. It is important to consider the unique features of the Arabic language and orthography and their corresponding design implications in promoting accessibility of Arabic electronic content. Although this research is in its early stages, we feel that our initial findings can give some guidance to practitioners on how to provide accessible Arabic content. Recognizing the diversity of individual differences between dyslexics when providing electronic content would have a substantial benefit for dyslexics and consequently for a much wider section of society. In the Arab world today, there is a great lack of awareness about Specific Learning Difficulties in general, and dyslexia in particular. More research in this area is required to better understand the needs of the dyslexic population so that they can acquire literacy skills and achieve their full potential.
References Abu-Rabia, S. (1998). Reading Arabic Texts: Effects of Text type, Reader Type, and Vowelization. Reading and Writing: An Interdisciplinary Journal, 10, 106-119. Abu-Rabia, S. (2001). The Role of Vowels in Reading Semitic Scripts: Data from Arabic and Hebrew. Reading and Writing: An Interdisciplinary Journal, 14, 39-59. Abu-Rabia S. (2004). Dyslexia in Arabic. In Smyth, I., Everatt, J. Salter, R. (Eds.), International Book of Dyslexia: A Cross-Language Comparison and Practice Guide (pp. 31-38). West Sussex: John Wiley & Sons. Abu-Rabia S., Share, D., and Mansour, M. (2003). Word Recognition and Basic Cognitive Processes Among Reading-Disabled and Normal Readers in Arabic. Reading and Writing: An Interdisciplinary Journal, 16, 423-442. Al-Sarhan, M. (2001). Report on Students with Specific Learning Difficulties. Ministry of Education (in Arabic). http://www.riyadhedu.gov.sa/alan/fntok/special/soobat.doc [date of access 27.12.05] Al-Sartawi, Z., Khashan, A., Al-Sartawi, A., Abujouda, W. (2001) Introduction to Specific Learning Difficulties. Riyadh: Special Education Academy Publications (in Arabic). BDA - British Dyslexia Association (2004). Dyslexia Style Guide. Available online at: http://www.bdadyslexia.org.uk/extra352.html [date of access 13.01.06] Beland, B., Mimouni, Z. (2001). Deep Dyslexia in the two languages of an Arabic/French Bilingual Patient. Cognition, 82, 77-126. Elbehari, G. (2004). Dyslexia in Egypt. In Smyth, I., Everatt, J. Salter, R. (Ed.), International Book of Dyslexia: A Guide to Practice and Resources (pp. 79-85). West Sussex: John Wiley & Sons. Evans, D., Blenkhorn, P. (2005). Designing a Screen Reader for Fully Sighted Users. HCI International 2005, Las Vegas, Nevada. Evett, L., Brown, D. (2005). Text formats and web design for visually impaired and dyslexic readers – Clear text for all. Interacting with computers, 17, 453-472. Goulandris, N. (2003). Dyslexia in Different Languages. Cross-Linguistic Comparisons. London: Whurr Publishers. Gregor, P., Dickinson, A., Macaffer, A., Andreasen, P. (2003). SeeWord – a Personal Word Processing Environment for Dyslexic Computer Users. British Journal of Educational Technology, Vol 34 No 3, 341-355. Holes, C. (1995) Modern Arabic: Structures, Functions and Varieties. New York: Longman Group Ltd. Kurniawan, S., Conroy, G. (2006). Comparing Comprehension Speed and Accuracy of Online Information in Students with and without Dyslexia. In Kurniawan, S. and Zaphiris, P. (Eds.), Advances in Universal Web Design and Evaluation: Research, Trends and Opportunities. Hershey: Idea Group Inc. Milo, T. (2002). Authentic Arabic: a Case Study. 20th International Unicode Conference. Washington, DC. Available online at: http://www.tradigital.de/specials/casestudies.htm [date of access 27.12.05] Rainger, P. (2003). A Dyslexic Perspective on e-Content Accessibility. Available online at: http://www.techdis.ac.uk/seven/papers/dyslexia.html [date of access 27.12.05]