Development of Writing Tools for Japanese Documents on a Personal Computer

日本的文字处理器设计成只用于输入、编辑、存储和输出文档，而不积极支持提高写作。我们已经开发了一个系统的写作工具，为机器可读的日文文件，称为SUIKESTA，它分析了日文文件的文本，并提供了作家的有用信息，抛光writitg。最初我们是在大型机上开发SUIKESTUDIO的。在本研究中，我们将SUIKtool移植到个人电脑上，扩充其功能，改善其使用者介面，并投入实际使用. SUIKtool运行在个人计算机NEC PC9801上，内存超过512KD。SUIKESTRUCTURE的程序是用C语言编写的。通过采用高速滚动和简单的多窗口，改进了用户界面. SUIKtool有两种类型的命令：原语和宏。后一种命令是通过组合原始命令为主要用户准备的。当用户调用时，SUIKtool会在屏幕上显示宏命令菜单。用户可以通过在菜单中选择功能来执行该功能。每个宏命令在一秒钟内响应，分析10000个日语字符的文档并显示结果。3.我们设计了原始命令，分别对句子集和字符串集进行操作。基本命令分为五组：显示、提取、排序、设置操作和杂项命令。为了实现这些抽取命令，我们设计了几种算法，这些算法不需要机器词典也不需要语法分析就能对文档进行文本分析，并从文档中抽取有用的信息，在实现这些算法时，我们使用了我们开发的适合日文文档的模式匹配技术.我们准备了一个在线手册，可以很容易地参考，而使用SUIKTOWN。

Japanese word processors are so designed as only for input, editing, stror, and output of documents, but do not positively support to improve the writing. We have developed a system of writing tools for machine-readable Japanese documents, called SUIKOU, which analyzes a Japanese document textually and provides the writer with the useful information for polishing the writitg. At first we developed SUIKOU on mainframe. In this research, we transferred SUIKOU on a personal computer, extended its functions, improved its user interfaces, and put it to the practical use.1. SUIKOU runs on a personal computer NEC PC9801 with more than 512KD RAM. The program of SUIKOU is written in the language C. By adopting the high speed scrolling and simple multi-window, we improved the user interfaces.2. SUIKOU has two kinds of commands : primitives and macros. The latter commands which are prepared for primary users by combining the primitive commands. When invoked by a user, SUIKOU displays the macro command menu on a screen. The user can perform a function by selecting it among the menu. Each macro command responds within a second to analyze a document with 10000 Japanese characters and display the result.3. We designed primitive commands to operate on sets of sentences and sets of character strings, respectively. primitive commands are classified into five groups : displaying, extracting, sorting, setoperating, and miscellaneous command.4. To implement the extracting commands, we devised several algorithms which analyze a document textually without machine dictionaries nor grammatical analyses and extract the useful information from it. While implementing these algorithms, we used pattern matching techniques which we developed to be suitable for Japanese documents.5. We prepare an online manual which can be easily referenced while using SUIKOU.