W3C home > Mailing lists > Public > www-math@w3.org > April 2003

Re: Content Markup or Presentation Markup for Audio Rendering of MathML

From: Neil Soiffer <neils@dessci.com>
Date: Tue, 29 Apr 2003 11:35:44 -0700
Message-ID: <3EAEC600.80400@dessci.com>
To: www-math@w3.org

Back at the beginning of April, I sent something similar to what follows to
this list and to Helder Ferreira, but it never made it to the list.
I'm resending the mail for the benefit other the people on the list.
I've amplified slightly what I said in earlier mail....

There have been a number of projects for spoken math.  I think that
most have focused on the needs of blind users, although there may
be other reasons (such as dyslexia).  In the later case, highlighting
the subexpression that is spoken is particularly important.

For examples of previous work, see:

MATHS (MathTalk) -- Robert Stevens, Allistar Edwards, York
http://www.cs.york.ac.uk/maths/
In particular, see
http://www-users.cs.york.ac.uk/~alistair/publications/ps/hci-journal.ps
for some experiments on audio rendering


ASTER -- T. V. Raman
http://www.cs.cornell.edu/Info/People/raman/aster/demo.html

MathSpeak (unambigious spoken math corresponding to Nemeth code) -- A. Nemth
http://www.rit.edu/~easi/easisem/talkmath.htm

MathTalk -- commercial product (mathtalk.com) which is part of a larger
bundle.

A. Karshmer (follow on to MAVIS project)
[the following is not free]

http://portal.acm.org/citation.cfm?id=274497.274524&coll=portal&dl=ACM&type=series&idx=SERIES368&part=series&WantType=Proceedings&title=ACM%20SIGCAPH%20Conference%20on%20Assistive%20Technologies
http://www.ercim.org/publication/Ercim_News/enw28/karshmer.html
Here's another, but much less useful reference:
http://www.ercim.org/publication/Ercim_News/enw28/karshmer.html

Karshmer did some experiments which proved what was propably intuitive:
spoken math is hard to remember for anything but small expressions.
For that reason, blind users need to use voice in conjunction with
expression navigation and braille displays or print.


To summarize the above and what I have found out:
1.  Prosidic and non-vocal information can aid comprehension.
2.  Spoken math is linear, math notation is not -- it is important for
     comprehension of all but very simple math that "navigation" is supported
     so that blind users can explore the structure of an expression
3.  Many blind users perfer to feel the math rather than hear it -- some
     form of Braille output (there are many different "standards" for
     Braille math) is important.  Both to embossing printers and to
     refreshable Braille displays.
4.  Low vision and people with some types of learning disabilities
     benefit from audio math.  These groups dwarf the size of the
     blind community.  Their needs differ.  Eg, synchronizing speech
     with highlighting can be very useful for some groups.  Larger
     fonts are useful for some groups.

Design Science is exploring accessibility issues.  We hope to make
MathPlayer accessible at some time in the future by incorporating
some of the above features and others into MathPlayer.

     Neil Soiffer
     Design Science
Received on Tuesday, 29 April 2003 14:42:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:54 GMT