On the unsuitability of horizontal scrolling

I have submitted this for publication so I am free to discuss it. Here is
the abstract. I can give a presentation this week.

This is a model built by a 69 year old mathematical who spent his entire
live using magnification to read. I used my visual experience and my
mathematical modeling experience to build this model. It shows that it
takes a lot more work to read with horizontal scrolling that it does to
read with word wrapping.

I think it establishes pretty conclusively that screen magnification
software alone cannot   provide accessibility support for low vision.

Please take time to read this abstract.

Wayne

*Abstract*

*Introduction:* The two methods to enlarge text for people with low vision
are: Magnification (Mag) implemented with lenses, CCTV systems and Screen
Magnification Systems (SMS), and Lexical Enlargement (LexE), text
enlargement with word wrapping that is implemented with software like word
processors and web browsers. LexE uses the lexical structure of language to
wrap words. This paper will show that it takes many more scroll actions
(keyboard and / or mouse) to read with Mag than LexE whenever the Mag
requires horizontal scrolling. Since all scrolling is overhead to the task
of reading, Mag introduces more overhead for reading.

*Method:* This study simulates reading using five text samples taken from
different genre. Models are built and analyzed using Microsoft Word and
PDF. Careful measurements are used to ensure a mapping to real monitor,
font and test column sizes. The author supplies a formula for computing
horizontal scrolling counts based on a grid model. The counting formulas
are valid whether the Mag technology is a hand-held lens, CCTV or SMS. LexE
scrolls were counted directly.

*Results:* Though costlier in scrolls than normal reading (27 to 1), LexE
is even effective at (2/3)-inch font when viewed on a 13-inch laptop. Mag
at 700% enlargement requires 98 times more scrolls than normal reading. At
moderate enlargement (300%, 400%, 500%), Mag requires more than five-times
the scrolling actions as LexE. It takes more scrolls using Mag at 300% with
horizontal scrolling than it does using LexE at 700%. This means, LexE
gives a greater effective range of enlargement than Mag.

*Conclusions: *The primary effort for providing accessible print and
digital content for people with low vision should shift from improving Mag
technologies to producing digital documents in formats that support LexE.