The art of computer programming, volume 3: (2nd ed.) sorting and searching
The art of computer programming, volume 3: (2nd ed.) sorting and searching
| Hi-index | 0.00 |
Bibliographic retrieval files frequently are indexed using words or phrases from a controlled vocabulary list. Usually these lists are structured to show relationships between terms, and strict rules are established to define permissible structuring. Because, with a large vocabulary it is humanly impossible to remember all of the many concepts involved, a printed list is required to define the different terms with their corresponding relationships. Computer processing is ideally suited for the term manipulation and tabulation required for generating such a list. Terms and relationships can be added or modified later, so that the vocabulary can remain useful for current ideas. Here again, computer programming can be utilized to make such changes. The Information Services Division of The University of Tulsa has used controlled vocabularies to index Petroleum Abstracts since 1965. Their Exploration and Production Thesaurus provides an example of the problems involved with managing a controlled vocabulary. A new Dictionary Control System has been developed for vocabulary maintenance, which is described in this paper. With this new system, thousands of terms with their corresponding relationships may be stored, modified or listed. A B-tree was chosen as the main data storage structure because of its multiway branching which decreases search time and storage requirements. A set of Dictionary System subroutines allows addition, deletion or modification of terms or relationships to be done interactively with individual terms, or in a batch mode. These routines allow extensive error checking using rules defined independently of the data structure. Experience with the new Dictionary Control System has shown it to be an efficient and valuable tool for controlled vocabulary maintenance.