Bert R. Boyce

189

Supporting Discovery in Virtual Libraries
Bipin C. Desai

Desai's work is a description of a problem and the outline of a solution that will be of interest to serious Internet users. Fragmented, incomplete, and less than compatible indexing tools for the Internet argue for a holistic approach to information discovery on the emerging information infrastructure. Current index generation systems are either manual, robot generated, or some combination of the two, and lead to highly variable results with the only consistency being a high number of items missed.

190

Data Set Isolation for Bibliometric Online Analyses of Research Publications: Fundamental Methodological Issues
Peter Ingwersen and Finn Hjortgaard Christensen

The metadata required for an adequate discovery system on the Internet include a standard index structure with standardized definitions and a facility for revision as sources change over time. A standard, easy-to-use software facility for information providers should allow provision of the needed information to distributed and replicated database servers. The Dublin Metadata Element List, DMEL, is described, and a graphical user interface for its application is suggested. Ingwersen and Christensen point out that for comprehensive bibliometric studies using online databases, cross file searching, database clustering, and duplicate removal are required features. It must also be possible to tune the retrieved set. The changing of the order of the files in the duplicate removal process will result in a different distribution of retained documents across the databases. One would desire to retain the most documents in files that had the desired searchable features (fields, search term breadth, phrase, or word by word indexing in the fields of interest, or name authority control). This likely means forming all overlap sets and examining them for desired characteristics prior to any ranking analysis.

205

End-User Searching Behavior in Information Retrieval: A Longitudinal Study
Weijing Yuan

Weijing Yuan investigates the effect of experience on the detailed use of search system protocols by law students using QUICKLAW over a period of 12 months. Mean number of commands and features used increased with experience. Use of phrases increased with experience, use of Boolean operators decreased as did the use of invalid commands. Search speed and use of field qualification seems to increase with experience. Attitude toward the system is unchanged with experience. No other pattern of change was detected. Dictionary and saved set features were rarely used. Confirming the useable terminology was uncommon, and building block strategies were rarely used. Truncation was used, but other word features were not. Large amounts of time were spent browsing and viewing results.

218

Ranking Schemes in Hybrid Boolean Systems: A New Approach
Jacques Savoy

Using average precision values over standard recall points, Savoy investigates the improvements that can be made by processing the retrieved set from a Boolean search using existing other than Boolean retrieval models to rank these results. There appear to be meaningful improvements, and in the process a good review of current retrieval models is provided.

235

A Discipline Independent Definition of Information
Robert M. Losee

Losee presents a general definition of information as being the values currently attached to characteristics in the output of a process. The output of a process is information about that process and its inputs and only in that context. The thought is that layers of communication processes occur, and that they accept input from other layers and produce output for other layers in a process called representation. Processes may be studied at any level, but movement between layers may result in lost information. The conceptual framework would appear to allow consideration of most existing ideas about information and could provide a commonality of approach.

254

Decision Support for Serials Deselection and Acquisition: A Case Study
David J. Robb and Angela McCormick

Robb and McCormick report a classic journal deselection project based on cost benefit analysis. Benefit is defined by a weighted sum of use, relevance, and availability factors collected locally. Of 670 journals, 206 were deselected but no study of faculty satisfaction is reported.

270

Citation Content Analysis of a Botany Journal
M. H. MacRoberts and B. R. MacRoberts

Replicating their earlier work in a journal on genetics which indicated that only 30% of influences evident in text are reflected in a paper's references, the text of an issue of Sida was studied by the MacRoberts to extract influences of previous work evident therein. Influences they judged present in the text appeared in the references only 29% of the time.

274

Electronic Publishing and Libraries. Planning for the Impact and Growth to 2003
by David J. Brown
reviewed by: Mary E. Brown

276

Civilizing Cyberspace: Policy, Power, and the Information Superhighway,
by Steven E. Miller
reviewed by: Donald Case

278

The Scholar's Courtesy: The Role of Acknowledgement in the Primary Communication Process
by Blaise Cronin
reviewed by: Betsy Van der Veer Martens

279

Digital Image and Audio Communications: Toward a Global Information Infrastructure
by Stanley N. Baron and Mark I. Krivocheev
reviewed by: Ruth Wuest

280

Design of Library Automation Systems: File Structure, Data Structures, and Tools
by Michael D. Cooper
reviewed by: Carol A. Hert

281

Explorations in Indexing and Abstracting: Pointing, Virtue, and Power,
by Brian C. O'Connor
reviewed by: Frank Exner, Little Bear

282

© 1998 , Association for Information Science
Last update: November 06, 1998