Non-BMP Character Support in Linux
Qian Xie - Institute of Software, Chinese Academy of Sciences

Intended Audience: Software Engineers, Systems Analysts, Testers

Session Level: Advanced

Due to increasingly demand of super ideographic character set, non-BMP Unicode character support will be crucial in Greater China Region. Considerable work has been done to survey non-BMP support status in Linux. The result is summarized and analyzed in the proposal and may be taken as a guideline of further development.

Non-BMP character supporting in a system needs each component performs properly and interacts to others consistently. To Linux, this is quiet difficult because of the diverse sources of components. In this paper, after a brief analysis of related mechanism, an application level survey of non-BMP character supporting ability is introduced and summarized. As expected, result shows there are pervading flaws in Linux desktop system. A bunch of test programs toward each critical library are designed to pin down the problem. Based on the test result, a list of underlying facilities with evaluation in term of non-BMP character support is provided. Several suggestions to application and toolkit developers are also made.

Toolkit and application developers may find this paper enlightening. Linux distribution vendors who aim at Greater China Region market will benefit as well. The providing information will help them to develop products that achieve Chinese user's demand.

This paper is for advanced audience. Attendees should have technical knowledge of both Unicode and Linux, especially of X window system.