"(Segletes) was able to apply a general-purpose linguistic parsing package he wrote called 'tokcycle' to solve the problem posed by the question,” Honeycutt said. “For example, if one were to use a screen reader to read the quadratic equation when the computer reads the equation aloud, it comes out ‘y equals a x two plus b x plus c.’ The computer reads ‘two’ instead of ‘squared’ because the computer must be told that the symbol means ‘squared’ before it can read the equation properly. By breaking math equations into proverbial phonemes using tokcycle, such as telling the program ‘Ok LaTeX, when you see 'x^5' in an equation, I want you to think of it as 'x to the fifth,' we can hide that ‘x to the fifth’ in the background of a PDF document.

Now when a standard screen reader sees this equation, it no longer says ‘x five’, it reads the hidden text of ‘x to the fifth’. This reads in a much more meaningful way, translating the codified hieroglyphics of mathematics into the understandable spoken linguistics...It is a big deal since LaTeX is such a core part of anything that renders math for viewing. This has the potential to give access and empowerment to more people.

And this is just a prototype....the only reason we have this tool now is because of the workshops that OU Libraries hosts. I was asked about this topic during a presentation on LaTeX, and was able to put the word out that people want this. Since the LaTeX community is a large, nebulous collection of software and typography nerds, no one thought about the need for screen reading improvements until they were informed about it. In that sense, OU Libraries facilitated the development of an accessibility tool." - Wesley Honeycutt