Letters to the Editor
SDI and 'Push Pull' Technology This letter was not published by The New York Times but has been incorporated into a more recent paper being submitted to
Journal of the American Society for Information Science
April 14, 1997
The New York Times
New York, NY
In her article on "... how I came to hate push technology," (The New York Times, p.C5, March 24, 1997) Denise Caruso speculates whether "Push Technology" signals the doom of the Web browser. However, on March 23, 1997 in "Pushy, Pushy," New York Times Magazine (p.32), James Gleick provides a cogent response. My experience with "Push Pull" technology may be of interest.
In 1965 Irving H. Sher and I created Research Alert, the first commercially available computer-based system for selective dissemination of information (SDI). Since then the service has been operated continuously on a weekly basis by the Institute for Scientific Information (ISI). The key to its success is timing, comprehensiveness, and high degree of specificity. Since the early seventies DIALOG, Lexis-Nexis, and other on-line systems have also provided "Push Pull" technology. The success of SDI services is based on their highly selective profiling systems. Unless "Push Technology" or current Web "crawlers" do the same, they will frustrate most users. Significantly improved search engines will make Web browsers increasingly valuable, even while equally improved SDI (Push) systems gain popularity.
Pointcast and other broadly based systems are relatively useless to most users but they can become highly specific, as they are with individual stocks.
The needs of scientists, medical researchers, and scholars are quite varied and only systems that can provide the ability to customize literature searching will be used repeatedly. Broader dissemination is provided by such tools as Current Contents and Medline, and hundreds of
leading specialty journals.
Profiling systems are widely used in the information industry to follow patent, journal, and other literature. The level of specificity needed often involves complex combinations of descriptors, but also the ability to identify current publications that quote specific papers and people.
Existing Web "crawlers" do not provide an acceptable level of precision and convenience, but competition will force them to rediscover what the library and information community has known for over three decades.
Eugene Garfield, Ph.D.
Institute for Scientific Information, Inc.
3501 Market Street
Philadelphia, PA 19104
cc : Denise Caruso
cc : James Gleick