Double-Byte Languages

Searching for information is sometimes doubly difficult.

English-language software is written in what's called single-byte code. In contrast, Chinese, Japanese, and other character-based software products are written in double-byte code. "Double-byte code, as the name suggests, consists of bytes that when put together make a single character, as used in many Asian languages," explains Dion Wiggins, director of strategic development for Asia On-Line, a Hong Kong-based Internet service provider (ISP) that specializes in furnishing content in Chinese, Japanese, Thai, and other Asian languages.

Double-byte software, however, is sometimes more expensive than single-byte pack ages. Plus, it poses problems in machine translation with single-byte-based syste ms, says Hong Chen, chief executive of Aimnet, which develops Internet software.

Adds Wiggins: "When searching English data [on the Internet], you can search for part of a word. But this is not possible with double-byte character sets as the characters themselves are the whole word."

Asia On-Line is developing a search engine based on neural network technology. "Neural networks learn from patterns and events, so when the same or even a similar event happens, it can make decisions based on previous events. If you apply this technology to a search facility, it can be used for any language, including all double-byte character sets and English," Wiggins says.