【作者】 芮建武；

【导师】 孙玉芳； 倪光南；

【作者基本信息】 中国科学院研究生院（软件研究所） ， 计算机软件与理论， 2005， 博士

【摘要】 全球化操作系统需要完善的国际化基础来支持对世界文字的输入、输出、存储与处理。本文从世界文字计算机处理的宏观角度出发,采用比较与分类、抽象与具体的研究方法,对操作系统框架、文字处理与图形用户界面等方面进行了研究,取得了六个方面的主要成果。 第一,深入剖析了基于POSIX标准的操作系统在各层次对文字的支持及其局限性,提出采用基于Unicode标准的操作系统实现框架。针对当前操作系统的实现现状以及与POSIX标准的兼容性,提出采用基于UTF-8编码的操作系统实现框架。这一工作解决了操作系统内部文本处理能力与文件系统输入/输出支持范围受限的问题。 第二,针对操作系统较弱的文字支持可扩展性、操作系统文字处理的不完整性与不一致性等问题,提出了基于管道—过滤器的国际化文字处理参考模型。模型比较了世界文字在文字体制方面的特点,总结出国际化文字处理的文字特征类别,提出了操作系统的文字处理逻辑,提出了以文字独立和文字相关模块为特征的软件结构,阐述了解决各类文字处理特征的技术途径,提出了国际化文字处理的标准过程。以此模型为基础,扩展了Qt系统支持库,构建了文字处理国际化基础。 第三,为了解决操作系统中面向文字的人机交互界面适应性问题,提出了基于部件的图形用户界面多文种适应性模型。模型分析了世界文字对图形用户界面的不同需求,抽象出适应各种文字的界面风格。通过把部件文字风格、部件的布局模式以及GUI区分为逻辑结构与交互结构两个抽象层次,实现了GUI的多文种适应性。根据此模型,扩展了Qt图形系统库,构建了图形用户界面的国际化基础。 第四,在上述模型的基础上,完成了传统蒙文操作系统的本地化设计。重点解决了蒙文字符的变形显现、蒙文文本的垂直书写和IS_TTB-LTR界面风格的GUI等与文字相关的处理。 第五,在上述模型的基础上,完成了藏文操作系统的本地化设计,此本地化系统完全基于ISO/IEC 10646标准。重点解决了藏文字符的叠加显现、藏文字符的变形以及梵音藏文的显示等问题。 第六,在上述模型的基础上,完成了维吾尔文操作系统的本地化设计。重点解决了维吾尔文字符的变形显现、维吾尔文文本的双向显示和IS__RTL-TTB界面风格的GUI等与文字相关的处理。 总之,本文在提出采用基于Unicode标准的操作系统框架、基于管道—过滤器的国际化文字处理参考模型和基于部件的GUI多文种适应性模型的基础上,完成了传更多还原

【Abstract】 Operating system provides internationalization services for applications, which generally indicates that to what extent internationalization of applications can be attained. A well built-in internationalization framework can be easily localized to adapt to all sorts of different local environments by means of separating culture-related or script-related data from common processes. Based on comparing and categorizing the orthographies of the world’s scripts, the most important parts of operating system internationalization have been working on in this dissertation and six achievements have been obtained.First, after analyzing the limitations derived from POSIX, an implementation framework of operating system based on Unicode standard has been described. This framework can extend the scope of processing multi-script text in file system and run-time libraries of operating system. Owing to the compatibility with POSIX, a solution utilizing the encoding of UTF-8 has been adopted.Second, an international text processing reference model based on pipe-filter paradigm has been presented, which aims to enhance the ability to support those new scripts more easily in operating system. This reference model consists of several parts, which include: the scope and content of text processing are clarified; the strategy on processing multi-script text is proposed; the structure of this model is proposed, which differentiates between general modules and those modules related to scripts; technical solutions are discussed to solve all sorts of different features of the world’s scripts; a standardized process of international text processing is summarized. Based on this reference model, Qt library has been extended to satisfy the requirement of international text processing.Third, a GUI multi-script adaptability model at the widget level has been proposed, which aims to meet the requirement of international GUI. Based on an analysis of writingstyles of the world’s scripts, different GUI styles are discussed. To achieve script-oriented adaptability of GUI, any interactive widget or layout object is considered as a combination of logical object, which is design-time object, and interactive object, which is run-time object. Thus by means of a graphical library conforming to this model it can be available to develop GUI of international application which is developed once and can run in any local environment related to certain script. An extension of Qt library conforming to this model has been fulfilled.Fourth, traditional Mongolian localized system has been designed. Because traditional Mongolian is written vertically from top to bottom in columns advancing from left to right and its letters always change their shapes depending on different conditions, it’s more difficult to implement a operating system supporting this script. There is no operating system fully supporting this script yet. Based on above models, designs related to this script are detailed, which include: displaying transformed forms of its letters, presenting Mongolian text vertically and adopting GUI with a style of IS_TTB-LTR.Fifth, Tibetan localized system has been designed, which conforms to ISO/IEC 10646. Tibetan text is spelled by overlapping its characters to form some vertically stacked characters, which is considered as a bottleneck to presenting Tibetan text. Within the framework of text processing reference model the most important issues are detailed, which include: displaying Tibetan stacked characters by a way of partitioning Tibetan syllables and applying an OpenType font, analyzing transformation of consonants and presenting Tibetan text transliterating Sanskrit text.Sixth, Uighur localized system has been designed. Uighur script used in China is written from right to left. Same letter may be written in different forms depending on how it joins with its neighbors. Some issues related to this script are discussed, which include: displaying transformed forms of Uighur characters, handling bidirectionality of Uighur script and adopting GUI with a style of IS.RTL更多还原