Journal of the Association for Information Science



Bert R. Boyce




 H.G. Wells's Idea of a World Brain: A Critical Reassessment 
W. Boyd Rayward

To begin this issue, Rayward examines Wells's concept of the "World Brain," provoking questions on the nature of today's global information systems. Wells advocated a New Republic, a world government growing from a world organization of scientific work and communication, an encyclopedia providing a systematic ordering of human thought and acting as a sort of superuniversity. As a physical organization it would involve collecting, summarizing, updating, and publishing the flow of new knowledge using microfilm and the information technologies of the day.

Wells also believed in man's evolution toward a  "conscious unification of the human species," by way of the superhuman apparatus of public knowledge toward social rather than individualistic goals. This involves eugenics, the breeding out of the less intelligent, the movement of scientists into the political process, and the deportation of criminal elements. The encyclopedia organization would speed this process by dominating worldwide education to create a common interpretation of reality. The Wellsian World Brain would function not only as a repository of scientific knowledge but as a database of the populace recording their characteristics and movements.




Literature-Based Discovery by Lexical Statistics
Robert K. Lindsay and Michael D. Gordon

Lindsay and Gordon explore a word count approach to Swanson's literature-based knowledge discovery strategy using complete MEDLINE records, two- and three-word phrases, and the identification of intermediary topics by high-occurrence frequency. Swanson's study linking migraine and dietary magnesium was duplicated. Ten of the 12 intermediate literatures previously found were identified.




Jumpstarting the Information Design for a Community Network
Misha W. Vaughan and Nancy Schwartz

Vaughan and Schwartz provide an example of a community service web site design based to a large degree on iterative user study information. Focus group sessions with paper prototypes and card sorts were used to solicit user opinion on how the site could differentiate itself from newspapers, libraries, and city government as a source, whether organization and labeling maximize meaningfulness to the user, and whether multiple categories might be reduced to a simplified hierarchy. Groups reduced the main categories to 10 to fit screen requirements and shared and discussed their results. All were asked to suggest services that might be provided different than those already available in the community. Discussion led to consensus on a structure, which was used to build a web site which was tested by eight participants each given 21 tasks to perform. A path followed by at least six was assumed to be appropriate. This resulted in the shifting, renaming, and cross-linking of several headings, and the removal of strongly community-oriented heading from an alphabetical display to the lead position.




Searching Scientific Information on the Internet: A Dutch Academic User Survey
Henk J. Voorbij

A random 1,000-person sample of the academic community of the Netherlands was surveyed by Voorbij's questionnaire. Of the 50% responding, 71% were Internet users. Students and faculty do not differ appreciably in levels of use. E-mail use is high, e-journal use is low. More traditional subject information sources rate above the Internet but it is heavily used to access factual and ephemeral material. Meta search and advanced search options are considered important but seldom used. Low precision, lack of quality sources, and response speed are seen as problems, but 68% believe results justify time invested. Lack of skill and of access are major reasons for non-use, but a significant number of non-users cited sufficient information elsewhere, and lack of knowledge as to what might be available in their disciplines. A focus group of 11 experienced faculty indicated a very positive attitude toward the World Wide Web, e-mail, and discussion groups. None were disposed to publishing on the Web.




SENTINEL: A Multiple Engine Information Retrieval and Visualization System
Kevin L. Fox, Ophir Frieder, Margaret M. Knepper, and Eric J. Snowberg

Fox et al. describe SENTINEL, a retrieval system using both an n-gram filter and a modification of the vector space model with vectors of documents judged relevant resubmitted in a feedback process, and documents ranked by combining their scores in both systems. Words with similar use are clustered together using a neural network training algorithm forming axes used for underlying positioning. The ranked list output with feedback capabilities is supplemented with a three-dimensional map of document and query positions based on the training set axes.





Systematic Weighting and Ranking: Cutting the Gordian Knot
Charles H. Davis and Geoffrey W. McKim

Davis and McKim describe the weighting and ranking algorithm of SWEAR™, which uses powers of two to assign weights to query terms entered. If N terms are entered, the first entered term is given a 2^N weight, the second a 2^(N-1) weight, and so forth until the last receives a 2^0 weight. Terms to be negated are given a negative weight equal to the sum of the positive weights, and possible terms a weight of 1 (2^0). A threshold is set at 2, and an accumulator for each document sums the weights of occurrences of query terms to generate a retrieval status value that provides a weak order of documents containing query terms. The searcher can change the weighting scheme, and thus the output ranking, by changing the order of entry.




Ink into Bits: A Web of Converging Media
by Charles T. Meadow
Reviewed by Jeff White



Technology and Privacy: The New Landscape
edited by Philip E. Agre and Marc Rotenberg
Reviewed by: Marc Lampson








asisnavbarASIS HomeSearch ASISMake A Comment

1999 , Association for Information Science
Last update: April 15, 1999