Once in a while, something nice happens, as if by coincidence, serendipitously. It happened to me when T.V. Raman asked me to supervise his Ph.D. thesis on building a system to speak documents, especially those with technical content or a lot of structure.
The project had many interesting points, for example: the need for a programming language for writing speaking rules (a sort of postscript for the ear instead of for paper), the need for a multi-dimensional model of speech and sound (so that the speaking-rule language could be largely independent of the particular voice synthesizer being used), knowledge of mathematics and the development of ways to speak it well, the development of a new internal form for mathematics, software engineering (in his 35,000-line CLOS program AS TE R, Raman makes ingenious use of object-oriented features of CLOS), and human-interface issues in making AS TE R interactive.
Finally, there was a real need for AS TE R. Previously, it was almost impossible for the visually handicapped to access technical documents. RFB (Recordings For The Blind) takes up to a year to make a recording of a technical book, and the only other way for a visually handicapped person to access a technical document is to have someone read it to them. The development of AS TE R is one more instance of the usefulness of computers.
Recordings For The Blind held a workshop-conference on AS TE R three months after the thesis was completed and will soon begin producing cassette tapes using it. Not only can a tape of a book be made in a day or two (instead of a year), but the quality is far superior to that produced by humans. In addition, Recordings For The Blind would like to make AS TE R available to the blind who are computer-literate, allowing them to make full use of AS TE R’s interactive capabilities.
The applications of the concepts and technology in AS TE R are many. For one, just consider being able to call your computer on the phone and have your e-mail read to you in a comprehensible fashion. The results are so noteworthy that SIAM News published a column on Raman and AS TE R (March 1994, page 7).
What makes this thesis even more intriguing is that Raman himself is visually handicapped. He cannot see his keyboard or monitor. He uses a guide-dog, Aster, to help him get around. And yet, Raman did all the programming of system AS TE R himself. My role was that of advisor. We discussed issues, I gave him advice based on my experience, I pointed him at literature, and I helped him in editing and organizing the thesis itself and a few papers that came from it (though all the writing on the computer was done by Raman himself). The real work, however, was his alone.
Raman is one of the most courageous, up-beat, and positive people I have known. Blind, he came from India to do his PhD work in Cornell’s field of applied mathematics. He acquired his guide dog, Aster, after arriving here. He studied applied math and computer science and began using computers in a big way (with a screen reader). All this without a complaint, with the most positive and optimistic attitude I have seen.
Just consider this anecdote. After his PhD, Raman accepted an offer at a major computer manufacturer. While still at Cornell, surfing the net, Raman came across a notice on a bulletin board to the following effect. Jack’s (a fictitious name) job was to go to T.V.’s future department and tell them how to treat Raman (because of his handicap). Jack, however, had little real experience with such matters, and he wanted advice from whomever would give it to him —what should he tell the department? Raman came into my room laughing about it. He had replied to Jack: tell the department to treat him just the way they would treat anyone else!
AS TE R is indeed a wonderful achievement. For me, however, AS TE R was not the most important part of this endeavor. Instead, it was working daily with Raman over a two-to-three year period. Always positive, with never a complaint, brilliant in his own way, Raman was a pleasure to deal with. It is an experience I will value for years to come.