Rollout!
========

Brian Dana Akers

Tetsuo stared at the pitcher of ice water. It was close to overflowing
and dripping with condensation. The ice sparkled under the halls bright
lights. It was so cold and inviting. He wanted to grab the handle--it
was only two feet in front of him--and rub the pitcher against his
graying temples, then empty it over his head. Of course, he could do no
such thing. Instead he reached out and sipped at his small glass of
water.

Tetsuo forced himself to stop daydreaming about the pitcher. He looked
around the convention hall. Three thousand people waited for his talk
and the all-important demonstration. The camera was set up in the center
aisle, ten rows back. He had been up all night supervising the sound
crew. Anything less than perfect sound would be a disaster. He caught
the eye of the sound man, who gave him a thumbs up sign. Tetsuo had been
up the night before last supervising the hardware technicians. A
hardware failure would be embarrassing and lose them marketing momentum,
but would not directly reflect on their new software. Three nights ago
he had been up with the software engineers, who had discovered a bug
that would have been absolutely fatal. All the problems had exhausted
Tetsuo. His career, his division, the company itself, their American
partner firm and even world affairs were on the line. The anticipation
for this closing event had been building throughout the three days of
the convention. Everyone was speculating about whether his team could
pull it off. He would be working in real time without a net, as the
Americans say.

The young man introducing Tetsuo was still droning away at the lectern.
Tetsuo was to have introduced his CEO, but the CEO had fallen gravely
ill last week and asked Tetsuo to take his place. Tetsuo had tried to
demur, reminding the CEO how shy he was, but the CEO insisted. Just look
at the three thousand in front of you and forget the three hundred
million watching through the camera, he joked. Tetsuo thought about
those three hundred million--all about to listen very carefully to his
every word--and started staring at the water pitcher again. It looked so
crisp and fresh and cool. He closed his eyes. He was imagining himself
skiing down a pristine mountain slope when he heard in the back of his
mind . . . and now the Vice-president of New Technologies, Tetsuo
Watanabe.

Everyone in the hall was applauding. Tetsuo stood up and carried his
glass to the lectern. He set the glass down and withdrew his pocket
prompter from his jackets inside pocket. He cleared his throat. Tetsuo
suddenly wished his English was a little better.

Thank you. Thank you very much, he began. It is a pleasure to return
to the heart of Silicon Valley on a sunny California day to be here with
all of you and give a small demonstration of our new product,
RosettaSoft. At any rate, one of his staff had said it was sunny when
he came back with the sandwiches. Tetsuo had not left the convention
center complex for the last three days.

I have many happy memories of coming to the Valley to work with our
American partners on this project. There were breakthroughs and
setbacks, progress and plateaus. Through it all, our sense of wonder at
the richness of human communication only deepened. But before I come to
RosettaSoft itself, allow me to briefly recap the highlights of almost
twenty years of development.

I was a newly minted computer scientist when I found myself--to this
day I still dont know how it happened--in a small discussion group that
included the editor in chief of our companys flagship publication. The
conversation had rolled around to world events. The editor remarked that
one inherent tension in the modern world was between homogenization and
diversity. A lively debate followed. Some held that the benefits of
world trade and communication were worth the cost of increased
homogenization. Others felt that material gain should not be placed
above ones cultural heritage. Still others felt you were fortunate just
to have the freedom to choose and not have one or the other forced upon
you. I mostly held my tongue, not being used to such lofty discussions.
Tetsuo took a sip of water and sneaked a glance at the audience. He
seemed to be holding everyones attention so far.

Gradually the conversation came around to this proposition: more than
dress, more than food, more than music; it is language that is most
fundamental to ones identity. What could be more dear to oneself than
the sweet sounds of the mother tongue? Tetsuo hoped that his language
was not too flowery.

It was at this point that the editor gave me a sly look and asked me
what I thought of automated translation. I replied that it was certainly
a knotty problem and that, decade after decade, the solution was always
just around the corner. He agreed that human communication was
profoundly complex and ambiguous. Then he added that a new standards
committee was forming and perhaps I should attend the first meeting.

Well, let me just say that my oldest daughter was in first grade when I
attended that first meeting and that I was still attending standards
committee meetings after she had graduated from college. Such exciting
meetings! The audience gave him a good laugh. He was relieved. Nothing
is more deadly than failed humor.

The biggest conceptual breakthrough for RosettaSoft came early on. We
had settled on the goal of universal translation software--software that
could translate text from any language to any language--and begun
research before we fully realized the magnitude of the undertaking. Even
if one leaves aside extinct writing systems and languages (and living
languages spoken by very small numbers of people) there
remaintwenty-seven actively used writing systems and hundreds of
languages.

Tetsuo started the first visual. The large screen above and behind him
showed seventy-five names of different languages scattered on it. Then
an animation began, inserting lines from each name to every other name.
Soon the lines blackened the entire screen.

As you can see, the number of language pairs is overwhelming-to say
nothing of the commercial viability of writing software to translate
Finnish into Zulu, for example. A small laugh this time.

We were sitting in a bar after work hours, dejected, when the solution
presented itself. Given our low spirits, we were quiet and
half-listening to the other patrons of the bar. There were
businesspeople from Germany, from Korea, from Malaysia, from Thailand.
Yet they werent speaking German, Korean, Malay or Thai. They were all
speaking English! It hit me like a thunderclap. I pulled out my pen and
took the napkin out from underneath my drink. The diagram I drew that
night in the bar looked essentially like this.

The second visual appeared on the screen. In the middle was the word
English, written very large, with the names of seventy-four other
languages arranged in a circle around it. This time the lines began to
appear like spokes in a wheel, connecting English at the hub to the
other languages on the rim. Very neat and tidy.

You see how it works? Just translate to and from English. Nothing else
required.

After this breakthrough, we made steady progress. We had a stable
product for the batch translation of text years ago. When my company
promoted me to the newly created position of Vice-president of New
Technologies, my first task was to prepare that product, code named
BabelBreaker, for its introduction.

Then that sly old editor in chief dropped by my office. He had retired
the year before and was using his time to read more widely and deeply
than hed been able to while working. He had been reading about the new
multigigachip technologies just then coming to market. Perhaps we should
add real time speech recognition and generation to BabelBreaker? What
people really want to do is communicate. What they want technology to do
is help--not hinder--them, he said.

I must confess my first reaction was not one of enthusiasm. I replied
that it would be a nontrivial undertaking. He said that he had already
chatted with the CEO about it--and the CEO was quite enthusiastic.

Back to the drawing board. Speech capabilities were added and we even
incorporated the ability to cooperate with third party recognizers and
generators. We retained the text capabilities and added the capacity to
generate transcripts from spoken input. And we renamed it RosettaSoft.

Does it work? Well, you can tell me. If you have understood what I have
been saying today--then it works. Two days ago, we made a fully
functional version of RosettaSoft available. Three hundred million
downloads and counting. Tetsuo swallowed hard and took a sip of water.
He did not look at the camera. Time to start wrapping up.

This product, this miraculous RosettaSoft those of you watching on the
Net are using this very moment, handles all twenty-seven actively used
writing systems and seventy-five languages. We will add ten more
languages by the end of the next calendar year. It truly marks the
beginning of a new era in global communications and--it is my fervent
desire--global understanding.

The audience rose to its feet applauding. The sound man gave him another
thumbs up. Tetsuo managed a smile for the first time in days. When the
applause ended, Tetsuo finished his presentation.

If I may remind you of just one thing: the version of RosettaSoft you
have is fully functional for today only. At midnight it turns into a
pumpkin! However, I am happy to say that our server is ready to receive
your orders for a permanent version. Given RosettaSofts remarkable
capabilities, we think you will agree that the price is quite modest.

To conclude our demonstration, I would like to introduce San Joses
very own International School Girls Choir. The audience applauded
again and Tetsuo returned to his seat. The lectern lowered itself and
disappeared. The house lights dimmed and spotlights lit up the stage.

A young Brazilian girl walked onto center stage and sang one line in
Portuguese. Then a young Indian girl dressed in a sari came out and sang
a line in Telugu. Then a Japanese girl dressed in a kimono, then a
Spanish girl, then a Tanzanian girl. One by one they came out and sang a
line until all had taken their place. They paused, then sang the chorus
of The Friendship Song in unison in English. They were an explosion of
song and color. The audience was thrilled and applauded loudly.

The house lights came back up and everyone began congratulating everyone
else. The demonstration had gone off without a glitch. Tetsuo patiently
answered the same questions over and over from members of the audience.
Finally he was able to leave the hall and go to his room in the
convention center hotel.

Tetsuo closed the door behind him and sat at his desk. Now came the
critical test. He turned on his computer and navigated over to the order
processing server. He entered his password and held his breath. Then he
saw the numbers. Orders were flowing in by the millions.

They had done it. It was a success. People around the world could now
access all the information on the Net in their own language.

Tetsuo stood up and stretched. He took off his jacket and loosened his
tie. Tetsuo felt very happy and fulfilled. He went into the bathroom and
splashed ice cold water on his face. Then he turned on his music player
and lay down for a well-deserved rest.

The forlorn sound of a shakuhachi grew faint in the distance, like the
moon setting in a dense fog. Daybreak arrived. A crane cried and took
flight.

Copyright 1997 Brian Dana Akers. All rights reserved.

