Crosswords are a popular recreational game that relies on the spatial relationship between words. As a player answers clues, they begin to organize words to form an intersecting grid. A good non-visual representation should convey the interrelation of words and support the user in building a practical spatial image of the crossword grid. This paper looks at two approaches to representing a crossword puzzle for visually impaired users: a screen reader based crossword, and an audio-tactile crossword puzzle. We evaluate the designs in a study with 10 visually impaired participants. The audio-tactile representation was found to support the practical use of the crossword's spatial structure while the screen reader based puzzle leveraged participant's prior experience in navigating websites. The paper discusses critical aspects of our study and presents a perspective on the use of multimodal interfaces for such spatial applications.
https://doi.org/10.1145/3313831.3376207
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2020.acm.org/)