A general purpose inverting indexing system

Abstract

The object of the thesis is to study the use of inverted File organization in a research environment. The main reason for the selection of this subject is the proliferation of information processing (3 systems, leading to duplication of programming and maintenance effort. Some other considerations to be taken into account include provision of Data Independence by use of Data Description Directories, together with User-Friendliness using tabular processing concepts based on the Relational Model of Data. A computer system has therefore been developed which embodies the above concepts in a practical system which is general purpose - i.e. it could be used to process any type of data provided it is presented in a tabular form. This system - called the General Purpose Inverted Indexing System – has been written in PL/1 and operates on the Northumbrian Universities multiple Access Computer (NUMAC) IBM System/360 Model 67 under control of the Michigan Terminal System (MIS). It comprises 24 modules totalling about 2000 PL/1 statements and includes the major functions required in a general purpose system, including Pile Description, Set-Up of Inverted Indexes and Processing of Queries. The thesis first describes inverted file organization concepts, and skin goes on to show how they are applied in a general purpose system. The components of the General Purpose Inverted Indexing System are then described, together with their interaction and operation. The logical and performance capabilities are defined and analysed; followed by recommendations as to how the system may be improved by future enhancements.