A crucial factor which affects the user-friendliness of a software system is the language used by the system to interact with its users. For this reason, software support for Chinese computing is becoming more and more important in Hong Kong as well as in other Chinese communities. However, most existing Chinese software systems are based on Chinese character sets that are fixed and small in size compared with the total number of Chinese characters, which has been increasing throughout the history. A solution used by most current Chinese systems is to reserve some user-defined code space and allow the user to define character glyphs for user-defined characters that are not in the standard character set supported by the system. However, this approach suffers from the encoding conflict p...[ Read more ]

A crucial factor which affects the user-friendliness of a software system is the language used by the system to interact with its users. For this reason, software support for Chinese computing is becoming more and more important in Hong Kong as well as in other Chinese communities. However, most existing Chinese software systems are based on Chinese character sets that are fixed and small in size compared with the total number of Chinese characters, which has been increasing throughout the history. A solution used by most current Chinese systems is to reserve some user-defined code space and allow the user to define character glyphs for user-defined characters that are not in the standard character set supported by the system. However, this approach suffers from the encoding conflict problem, i.e., the same user-defined code may represent two different characters in two different systems. Problems arise when data exchange between systems is necessary.

A novel approach to Chinese character encoding is proposed in this thesis. This approach is based on escape sequences and the composition of user-defined characters from their structural components. With this new approach, the user is no longer required to assign reserved codes to user-defined characters. Instead, expressions representing the composition of characters from their structural components are specified. Hence, the encoding conflict problem can be avoided.

To solve the problem of generating character glyphs from character composition expressions, some traditional Chinese calligraphy rules are studied. These relatively abstract and qualitative rules are quantified in this thesis to define some Chinese character beauty evaluation metrics, which are then used to guide the automatic generation of glyphs for display through a greedy search process.

A prototype system has been implemented to demonstrate the ideas proposed.