Dissertation Information for James Harvey May NAME: - James Harvey May
DEGREE:
- Ph.D.
DISCIPLINE:
- Library and Information Science
SCHOOL:
- Columbia University (USA) (1978)
ADVISORS: - Theodore C. Hines
COMMITTEE MEMBERS: - None
MPACT Status: Incomplete - Ambiguous
Title: IDENTIFICATION OF DESIGN PRINCIPLES FOR EFFICIENT COMPUTERIZED BIBLIOGRAPHIC DATA INPUT
Abstract: One of the most important aspects of information systems from the standpoint of cost is the input process. A critical part of that process is the keyboarding of data. It is the costliest single operation in the conversion of data into machine-readable form for major bibliographic data production systems today. Cost of keyboarding can be reduced through attention to frequency and ordering characteristics of specific items of keyboarded data. These characteristics are determined from an analysis of present large scale systems, especially MARC II of the Library of Congress. A variety of statistical data in printed, data base, or unpublished manuscript form has been gathered on some of the more important of these systems. That data has been supplemented by analyses of frequency and ordering compilations found elsewhere. The various statistics are shown to be interrelated and consistent over a variety of situations. Bibliometric laws are seen as useful in predicting frequency patterns of character strings. Actual techniques to reduce input currently used by the systems studied are reviewed. Orders of magnitude of their savings in keying are determined, where possible, through reference to data characteristics. These techniques are of two types depending on the kind of data characteristics they use. One type relies on how relative frequencies of particular symbols, strings, words, phrases, and bibliographic elements are distributed in specific instances. Basically, more frequently occurring strings are made shorter, in terms of keystrokes used, through various abbreviating schemes. With limited size keyboards rarely occurring symbols, on the other hand, may be converted to character strings so that keys are freed for frequent symbols. The second class of techniques takes advantage of source data patterns, or order. Bibliographic data is particularly amenahle to this class of techniques because of its consistent formats and the hierarchical structuring of much of that data. The internal pattern of an entry, in terms of the bibliographic elements selected and the order in which they are found on the source document, is fairly rigidly set and, usually, can be predicted, especially for catalog entries. Entries such as journal article citations are grouped by issue so that data cornmon to a series of entries can be keyboarded only once and automatically posted to each entry when the data is processed by the computer. Alphabetic or other order of whole files to be converted to machine-readable form is also shown to be of use in reducing input keying. From the review of techniques several general principles of input design are found. These general principles suggest other ways to make current techniques more efficient. The principles also indicate other ways of dealing with input. A subset of possible improvements in input design which is relatively machine and data dependent is chosen for analysis. The principles are tested on actual large and well-known systems. How these principles fit into the broader theoretical bases of information systems is also explored. In terms of semiotics, or the study of signs, the improvements analyzed utilize syntactic rather than semantic or pragmatic relationships. Yet, underlying principles are shown to be conducive to expanded application with the higher semiotic levels. Ideas for further research in these areas are given.
| |
MPACT Scores for James Harvey May A = 0
C = 0
A+C = 0
T = 0
G = 0
W = 0
TD = 0
TA = 0
calculated 2008-01-31 06:01:51
Advisors and Advisees Graph
generating graph, please reload |