|Advertising|Jobs 転職|Shukan ST|JT Weekly|Book Club|JT Women|Study in Japan|Times Coupon|Subscribe 新聞購読申込|
|Home > Life in Japan > Features|
|Home > Life in Japan > Features|
Friday, Nov. 15, 2002
THE ZEIT GIST
Putting Japan's first bilingual WP to the test
Mixing English and Japanese was no easy task
Second of two parts In my previous installment, I noted that Toshiba launched the first dedicated Japanese-language word processor in 1979. Five years later, the Japan subsidiary of MicroPro International Corporation, publisher of WordStar, the pioneering English word processing software, was preparing to launch "WordStar 2000" for the IBM PC.
MicroPro's strategy was to port this new program, code-named "Kanji Ivan," to run on Japanese PCs as the first bilingual Japanese-English software.
Excited to have a chance to work in this promising new sector, I joined MicroPro Japan in April 1984.
Why was bilingual software necessary? Dedicated Japanese word processors worked fine in Japanese, but did a pretty awful job with the Roman alphabet.
Letters were handled as two-byte characters, i.e. no differently from kana or kanji, and offered no choice of fonts or even simple enhancements like italics.
That meant that when Japanese and English were mixed in the same document, the result, from an aesthetic standpoint, looked terrible.
For one thing, "word wrap," the glue that holds English words together each time you reach the right margin, was not applied, so you often winced at "orphans," one or two letters by themselves on the far right margin, followed by the remainder of the word on the next line. Automatic hyphenation? Spell check? Thesaurus? Computers still lacked the sophistication to incorporate those tools into a Japanese program.
MicroPro not only aimed at offering fully functional English-language word processing, but also mixing Japanese and English in the same document -- even in the same sentence -- while sharing virtually all the same editing and formatting commands, such as underline, italics and so on. Such software had a ready market for academics, translators and others who worked in two languages.
We held a name-the-new-baby contest and one of MicroPro's programmers, Kaoru Abe, coined the name "TwinStar." I was transferred from documentation to work as a tester on the program's pre-release version, and I spent eight to 10 hours a day putting it through its paces.
One testing method was to enter impossible commands and watch to see whether the computer displayed the appropriate error message or crashed.
Our team of "code-busters" passed along a steady stream of "bug reports" to Abe-san and the other programmers, who would duplicate them and then put a "patch" on the code.
To do my job, I had to learn how to type in Japanese. We did not have a manual because our software was still in its beta stage of development, so my education was mostly through trial and error.
The keyboard on our NEC 9801 PCs had, give or take, 108 keys. But the Japanese written language has 96 phonetic characters and thousands upon thousands of kanji, not to mention tens of thousands of compound words.
How, then, was I able to get them up on the screen and print them out?
As I mentioned previously, Toshiba had paved the way for Japanese word processing by developing intelligent algorithms to process the keyboard input in two stages.
First the phonetic equivalent of the character goes to the internal dictionary, looks up the candidates, and displays them on the screen.
The user then makes a selection from the screen and transfers it to the document.
Of course, if you type the wrong keys, you get the wrong kanji.
What is really frustrating, though, is that in order to input Japanese into a word processor, I basically had to relearn how to "spell" the Japanese language.
The box above shows just a few examples on the rules for "spelling" Japanese on a word processor.
Nearly 17 years have passed since I worked on the development of TwinStar. While PCs like my own Mac G4 have become vastly faster and smarter, inputting Japanese via the standard QWERTY keyboard, alas, still remains an extremely cumbersome task.
Is the situation hopeless?
Far from it: Thanks to its comparatively simple phonetics, Japanese is ideally suited for a system based on voice recognition.
It may be just a matter of time before keyboards are rendered obsolete and people input words into their computer in real time -- by talking to them.