How Much Effort is Necessary for Blind Users to Read Web-based Mathematical Formulae?

Abstract

Providing access to mathematical content on websites for blind users who use screen readers has been an important challenge for accessibility in Human-Computer Interaction. Further to limitations in reading content in languages such as Brazilian Portuguese, there are many issues related to the effort demanded from blind users to navigate within mathematical formulae when solving problems to reach for specific information. In this paper, we performed estimations of the effort required by blind users to navigate within mathematical formulae using task models with GOMS (Goals, Operators, Methods, and Selection rules) and KLM (Keystroke-Level Model). We analyzed task models for different mathematical formulae using available navigation resources in the screen readers JAWS (Job Access With Speech), ChromeVox and NVDA (NonVisual Desktop Access), considering the optimal paths that could be used by experienced blind users. The results showed that JAWS currently provides more detailed navigation resources that allow for faster and more precise navigation than ChromeVox and NVDA. However, we observed that there are important issues regarding the complexity of the interactions required to use those navigation resources. This could be very challenging for blind users when performing mathematical problem solving tasks, which are cognitively-demanding themselves. The investigation of new enhanced interaction approaches for within-formulae navigation in screen readers is very important to make mathematical learning and problem solving on the Web easier for blind users.