Blind computer users can use the computer to produce documents using screen readers and speech synthesizers. The disadvantage is that common mistakes such as inconsistent capitalisation in headers and missing full stops are not apparent when using a speech reader unless each individual character is spoken which is tedious and time consuming. The aim of this project is to produce a tool that will speed up the correction process by identifying errors missed by conventional spelling and grammar checkers.
The end product will be similar to that of the C programming tool Lint which picks up mistakes missed by the compiler.
A number of samples of computer produced text from a blind writer were studied to identify the type of errors produced. The text was in the form of LaTeX which produced extra errors to those in normal text, in particular mis-matched brackets and using reserved characters without an escape code. There are several LaTeX checkers available including ChkTeX. Errors picked up by ChkTeX were not replicated in the implmentation.
From these errors a suitable software implentation tool was chosen, in this case Perl. Perl was ideal, due to its speed, comprehensive inbuilt string handling functions and cross platform availability under the GNU license scheme.
Finally, the tool will be tested against the original samples and a critical evaluation of its performance will be made along with suggestions for improvements and further investigation.
A particular area of study would be for the interface to the tool to achieve maximum ease of use and benefit for a blind user.