Frank Bokhorst wrote:
> 
> I am looking for help in dealing with text that contains mathematical
> symbols, which has to be recognized by a text-to-voice conversion
> system. We are using Jaws, in this case.  This is for the benefit
> of two blind students I have in the class, and who are bravely
> attempting a major in Psychology. The text is D. Howell's
> "Fundamental Statistics for the Behavioural Sciences".
> 
> First we scan the text using Omnipage into an electronic version.
> However, Omnipage skips over math symbols, and we have to edit the
> output to make it suitable for Jaws, which converts it to voice.
> When dealing with an entire textbook, this can be very slow work.
> 
> This university has a well-resourced Disability Unit that our students
> can rely on, but my course textbook remains a problem for them.
> 
> It would be nice to hear from other people with experience in this
> kind of process. Perhaps someone has useful suggestions for making
> it more efficient, or persons to make contact with?
> 
> Other than the Howell text, are there perhaps more general materials
> available for teaching basic applied statistics with blind students?

        I do NOT have experience in this area, but here's my $2E-02 anyway.

        I would, were I in this situation, probably approach the author and beg
a copy of the word processor files from which the text was generated -
offering in return to send the author and publisher a copy of the final
product. 

 Simple math formulae may well be there just as symbols from a character
font; this gets around OCR problems.  It should then not be impossible,
using a Visual Basic script, to scan specifically for characters in such
a font and replace them with phonetic equivalents.

        If by any chance TeX had been used a simple search-and-replace would
handle many inline formulae, leaving a handful that may need to be
explained by hand.  This would not work for more avanced formulae, but
any thisn "fundamental" and "for behavioural sciences" ought to be
avoiding anything too hairy anyway.

        -Robert Dawson
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to