BIOINFORMATICS - SEARCHING THE NET

During the past 30 years, there has been an explosion in the volume of published medical information. As this volume has increased, so has t he need for Efficient methods for searching the data. MEDLINE, the pri mary medical database, is currently limited to abstracts of the medica l literature. MEDLINE searches use AND/OR/NOT logical searching for ke ywords that have been assigned to each article and for textwords inclu ded in article abstracts. Recently, the complete text of some scientif ic journals, including figures and tables, has become accessible elect ronically. Keyword and textword searches can provide an overwhelming n umber of results. Search engines that use phrase searching, or searche s that limit the number of words between two finds, improve the precis ion of search engines. The development of the Internet as a vehicle fo r worldwide communication, and the emergence of the World Wide Web (WW W) as a common vehicle for communication have made instantaneous acces s to much of the entire body of medical information an exciting possib ility. There is more than one way to search the WWW for information. A t the present time, two broad strategies have emerged for cataloging t he WWW: directories and search engines. These allow more efficient sea rching of the WWW. Directories catalog WWW information by creating cat egories and subcategories of information and then publishing pointers to information within the category listings. Directories are analogous to yellow pages of the phone book. Search engines make no attempt to categorize information. They automatically scour the WWW looking for w ords and then automatically create an index of those words. When a spe cific search engine is used, its index is searched for a particular wo rd. Usually, search engines are nonspecific and produce voluminous res ults. Use of AND/OR/NOT and ''near'' and ''adjacent'' search refinemen ts greatly improve the results of a search. Search engines that limit their scope to specific sites, and metasearch sites that use multiple search engines optimized for specific types of searches have recently emerged. The distinctions between search engines and directory searche s have blurred. Eventually, conceptual searching in which the computer searches for related ideas, without having; to be given all the relat ed keywords, may become a reality. This will free the user from having to learn specific rules about searching, allowing energies to be focu sed on results of the search, not the search itself. Copyright (C) 199 8 by W.B. Saunders Company.