T.V. Raman is grateful that he didn't completely lose his
eyesight until he was 13. Because of that, he says, he didn't get
used to always receiving ``special'' treatment.
``When you get things done special for you, you come to always
expect it,'' said Raman, who was blinded by childhood glaucoma. ``It
can be a crutch. You get marginalized. The sooner you become
mainstreamed, the better.''
That's the same attitude Raman brings to his research in speech
technology. While Raman, a software engineer, is a fervent proponent
of making the Web and other technology accessible to the disabled
community, he chooses not to work in the assistive technology field
where products are designed specifically for people who are
physically impaired. Instead, he works in developing mainstream
products, such as cell phones and personal digital assistants that
can be used with equal ease by people with disabilities and those
without.
``With most assistive technologies, you end up retrofitting
something that hasn't been properly designed to work for people with
disabilities,'' said Raman, a researcher in the Human Language
Technologies Lab at IBM's Almaden research facility in South San
Jose. ``You want to design it right from the beginning and make it
work for you.''
Raman, 38, is one of a small community of engineers creating new
technologies that will make the Web much more responsive to the
needs of all users. The standards that he and other researchers are
working on will form the framework for the next generation of Web
applications. Raman's goal is to allow Web developers to create a
single piece of content that can take many different forms -- text,
audio and video -- to accommodate many different types of users.
Accessibility
Step into Raman's spartan office and one striking detail jumps
out -- there are no papers or books anywhere. ``There's nothing I
can do with paper,'' laughs Raman. Nothing, that is, except tear it
up and fold it into intricate designs, which he does when he
fidgets. Raman, who sports Ray-Ban sunglasses to shield his cloudy
white eyes, empties his pockets for a visitor and they are full of
oragami-like structures.
Raman's Web research is done in conjunction with the World Wide
Web Consortium (W3C), the international standards-setting body that
works to ensure the Web develops in a cohesive and relatively
seamless manner. The group, headed by the Web's original creator,
Tim Berners-Lee, is on the forefront of efforts to promote
accessibility for the disabled.
When the Internet was young, and primarily text-based, blind
users could access most of it through text-to-speech software. As
the Net evolved and began adding bells and whistles -- graphics,
audio and video -- it became less accessible.
That evolution has forced developers to find new ways to bring
information to disabled users, said Jim Fruchterman, who is working
with W3C on accessibility issues and helped draft some of the
federal government's Web accessibility regulations in 1999.
``Before, the goal was separate and equal. But of course once
it's separate, it very quickly stops being equal,'' said
Fruchterman, who runs Benetech, a non-profit technology company in
Palo Alto. ``Now, the goal is universal design that's integrated and
equal: Don't make disabled people use a different Web structure but
make it so they can use it too.''
Raman's research for IBM, which overlaps with his
standards-setting work for W3C, includes developing speech
technology for a variety of devices. He's also working on a
technology called ``x-forms,'' which would make it easier to collect
data on the Web. For example, a Web form that now must be completed
by typing could be filled out by voice or by a beamed message from a
PDA.
The myriad technical challenges of making the Web a more
responsive medium are just another set of puzzles to someone who
devours all manner of math games and brain teasers.
Raman's eyes began to deteriorate during his childhood near
Mumbai, India. Because the slightest fall or injury could trigger
further damage to his eyes, he spent most of his childhood as a
bookworm in the corner of the room while his cousins and other
children ran and played around him. His love of puzzles and numbers
came early. Even after he was fully blind, he could twist and turn a
Rubik's Cube (specially outfitted so his fingers could distinguish
the textures) in seconds. These days, most of the the counter space
in the San Jose condominium he shares with his Labrador guide dog,
Hubbell, is taken up by puzzles and elaborate geometric structures
made of children's building toys.
No more gibberish
After graduating with a bachelor's degree in mathematics and a
master's degree in computer science in India, Raman came to Cornell
University to study applied mathematics. It was then that he began
using speech software. In his applied mathematics class, the
software read back gibberish when it tried to read complex
mathematical equations on the screen. So Raman designed his own
software that could decipher the formulas. That project, which
eventually turned into his doctoral thesis, launched his interest in
speech technology.
In 1994, Raman wrote a program, emacspeak, that allows his
computer desktop to be completely audio-enabled. A computer voice
reads to him what's displayed on his screen as he navigates e-mail,
surfs the Web and accesses programs. Raman's software is distributed
free on the Web and IBM will package it soon as part of a suite of
server software. It's also distributed on Sun's operating system,
Solaris.
On this particular day, Raman is showing visitors how the program
works. He opens his e-mail program using his keyboard and the
computer voice reads the messages to him at warp speed. It sounds
exactly like playing an audio tape on super-fast forward. With
nothing to see, all of his concentration goes to his ears. ``That
way, I'm hearing it as fast as you would read with your eyes,'' he
said.
Raman joined IBM in 1999 after having worked for Digital
Equipment and Adobe.
He gets excited when he talks about XML, or Extensible Markup
Language, which Raman and others believe is a linchpin in making the
Web more accessible.
Specialized needs
The problem with the Web these days, Raman and others say, is
that content takes a back seat to presentation. And snazzy Web pages
with lots of graphics, streaming audio and video content are often
useless to people with disabilities.
XML is a system of encoding Web content so programs can work
together more intelligently. Once content is written in XML, it can
be displayed as plain text, audio or graphics.
``The separation of form from content is a key tool for
accessibility, but it's also a good design technique for lots of
other reasons,'' said Berners-Lee. XML ``helps one adapt information
to different applications, needs, devices, cultures, languages, and
so on.''
Advocates of XML fervently believe its adoption will not just
improve access for the disabled but lead to better products and
technology for everyone.
As devices such as cell phones get smaller and smaller and the
world becomes more wired, ``who's going to want to read the
newspaper visually by looking at that small screen on your cell
phone?'' asked Raman. ``You're going to want your phone to read you
the newspaper.''