Name
Abstract | ||
|---|---|---|
This paper describes a file system which is used widely in Massey-Ferguson to implement nontrivial files. It has been used for data bases of all sizes, the largest of which are several megabytes. As a result of using this file system in all our major systems we have found several benefits, including: - - new databases can be designed and implemented very rapidly indeed - often in a matter of hours - using standard keyed file functions. - The design of the keyed file usually relates closely to the “real world”, making the system very transparent and maintainable. This contrasts with complex indexing structures which - while they may indeed be more efficient than keyed files - usually make the database incomprehensible to anyone who has not studied it at length. Keyed files are simple to use, understand and document. The structure of one of our largest databases can be easily described in a few sentences, and understood immediately by anyone familiar with the keyed file system. - We can quite easily make our systems communicate with each other (at the technical level). The paper begins with an outline of the Sharp APL File Subsystem, on which keyed files are based. It then describes the keyed files approach, including the design objectives. The conceptual structure of a keyed file is illustrated. All keyed files - whatever their particular manifestation - have this structure. The next section discusses some of the considerations in designing a keyed file, and indicates some of the limitations of the system. The paper closes with an example of an application, describing how the database is structured to face the problems it is used to solve. The author has worked for Massey-Ferguson since 1961 and first became involved with computers in 1963. He has programmed in many languages for several different computers. He was the pioneer of APL when it was introduced to the Corporate Head Office of MF in 1972. The language is now used for almost all Corporate systems, and for many planning and monitoring systems in other locations. These systems range from long-range planning models to daily monitoring systems to special one-off analyses, and vary in size from a few lines to several thousand lines of APL code. |
Year | DOI | Venue |
|---|---|---|
1976 | 10.1145/800114.803668 | APL '76 Proceedings of the eighth international conference on APL |
Keywords | DocType | Citations |
apl code,file system,corporate system,sharp apl,standard keyed file function,major system,daily monitoring system,keyed file,nontrivial file,keyed file system,keyed files approach,measurement | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 1 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| M. Elliott | 1 | 0 | 0.34 |