So many Chinese, but too many Chinese characters

One billion people live in mainland China, but will they ever see a computer? That question may be answered on Taiwan, where IBM Corporation, Wang Labs, and other companies are battling over which of their various computer setups best fit the Chinese language.

Compared with the 26 letters of English, the number of Chinese characters complicates the task of even the largest computer memory. The Chinese Oxford Dictionary lists some 35,000 basic characters, with upwards of 300,000 possible variations.

IBM's new 3270 "Chinese Information System" uses 11,560 characters, with a smaller set of frequently used words. The first one was installed this year for the Taipei provincial government.

The stakes are high: The market is large in Taiwan, with banks and other computer-hungry institutions needing to transmit data to one another. And as a future opportunity, mainland China could be a lucrative market.

The Taiwan government believes it can develop an "intercode" program that would enable the dozen or so variations to talk with one another. But IBM officials think that won't work. "Chinese is an open-ended language, which allows for creation of new words," explains Paul S. Chan, IBM's manager for Taiwan. "The government needs to designate a set of characters. Then you box the problem. The Japanese did it 10 years ago to develop the computer industry, and we will have to do the same."

An intercode language "would be like two aliens talking through a third language," Mr. Chan says. "Why do it at all?"