摘要： 随着人类社会信息化程度的提高,文本数据库的应用会迅速增长。但是,由于文本数据库包含的信息具有非常自然的特性,对它的查找和检索是相当复杂的。传统的冯·诺依曼机器仅适用于对数值进行运算,而计算机情报检索所执行的操作,如模式匹配、查找、排序或检索等。都是对存贮的内容进行判断。所以尽管目前有许多规模大小不一的文本情报检索系统在运行,但是赖以这种传统硬件结构的文本检索软件系统的实际效率却是十分低下的。为此要开发适合于非数值运算的文本检索的硬件结构。本文从系统一级描述了几种文本检索的后端处理机结构和配置,这些后端处理机都有能力完成整个情报检索过程,或至少是这种过程的一部分。最后,从部件一级描述了几种用于文本检索的具有非数值运算特定功能的硬件部件。

Abstract: The Application of textual databases is on the increase with the rising of informatio-nalized level of human society. However, search and retrieval operations on textual databases are complicated by the very natural characteristics of information they contain. The actual efficiency of text retrieval software system based on the conventional hardware architecture is very low. It is necessary to develop a hardware system suitable for non-numeric operations of text retrieval. This paper describes several kinds of backend processors, which have the capability of completing the entire proceeding of information retrieval, and hardware parts, which possess non-numeric operation function.