|Home > Life in Japan > Features|
|Home > Life in Japan > Features|
Friday, Nov. 1, 2002
THE ZEIT GIST
Japan goes from clunky typewriter to waapuro
The nuts and bolts of Japanese typing
I wonder how many readers have ever experienced typing on an old-style Japanese typewriter. I tried my hand at it, just once. It was around 1973, and afterwards I was relieved that my clumsy effort was merely done out of curiosity and not necessity.
Seen mostly in large companies and government offices, the old typewriters were typically operated by a specially trained typist, usually female. "Personal printing press" would be a more apt description for these remarkably cumbersome devices than "typewriter."
"Typing," in the western sense, fails to acknowledge the expenditure of effort needed back then to get Japanese onto a sheet of paper. The typist manipulated a round white knob over a flat tray as she searched for, selected and physically extracted rectangular metal "slugs" bearing a reversed image of the hiragana, katakana or kanji character from the tray. Then she depressed a lever to slam the slug against the ribbon, print the character on the page, and replace the slug back in its original space.
Instead of the the stacatto rhythm of a touch typist rapidly pounding on keys, one heard faint whirring noises, followed by "WHACK! (pause) WHACK! WHACK-WHACK! WHACK! (longer pause while she searched for an unfamiliar character) WHACK! (pause) WHACK!
Credit for making the old typewriter obsolete goes to Toshiba, the company that developed Japan's first word processor. This story was recently told in an English-language book from Penguin published earlier this year entitled, "TOSHIBA: Defining a New Tomorrow," by Robert L. Cutts.
Well before it built the word processor, Toshiba's engineers set out to develop a scanner to help the post office automatically sort handwritten addresses on letters. A team headed by Dr. Kenichi Mori began its research in 1962, and five years later delivered the finished product to Tokyo's Central Post Office.
With the project completed and his team reduced to a handful of people, Mori looked around for another quest and decided on machine translation.
"We knew we had to develop a keyboard that could somehow input the characters of Japanese and all their subtle meanings in accord with their usage in the syntax of the actual text," Mori recalls. "Plus we had to give it a big memory -- and we had to make it portable."
The challenges of teaching a computer to type Japanese were immense. When typing on standard 108-character keyboards with the QUERTY configuration, which most typists use, the key you strike on the keyboard doesn't correspond to what eventually appears on the computer screen.
While the keyboard commands may differ slightly depending on the machine and software in use, word processor input in Japanese works as shown in Fig. 1 (above).
Let's say you want to type momo, meaning "peach."
When you type the letter "m," it stays on the screen until the vowel "o" is typed; then it automatically changes to the hiragana "mo." But momo, in addition to "peach," can also mean "thigh" (written using two different kanji) and "hundred." So the typist must make a selection from those four. The syllable "kan" has 60 or 70 possibilities; the name "Akira" can be written about 25 different ways.
Compound words also present an even more difficult challenge, because Japanese has lots of homonyms. As Cutts' book explains, you can't just type in "kankei" and see the word for "connection" on your screen. You have to tell the machine that you mean "connection," as opposed to several other kankei that mean the return of the Crown Prince, a plumbing diagram, an iniquitous plot, severe punishment, etc.
To deal with this complexity, Mori's team at Toshiba developed intelligent software, so that when the typist enters kankei, the screen will display a full vocabulary list of all words pronounced that way, in order based on their frequency of usage. An example of this is shown in Fig. 2.
Toshiba launched its first dedicated Japanese-language word processor in 1979. The system sold for a mind-boggling 30 million yen -- more than 100 times what you can expect to pay today for a good PC, software and a printer.
For his research and development of the Japanese word processor, Dr. Mori was honored with the Minister of State for Science & Technology Award by the government.
The first dedicated Japanese word processors like Toshiba's worked fine when typing only in Japanese, but did a poor job with the Roman alphabet.
In 1984, the Japan subsidiary of MicroPro International, publisher of WordStar, began an ambitious project to develop the first bilingual wordprocessing program for MS-DOS computers.
The next time, I'll be writing about my own involvement in this project.