Left ArrowReturn to the main page

The Search Problem

The search problem. There are millions of Web pages, no one knows how many, it's like trying to estimate the number of pebbles on the beach. A number of services exist which allow users to search the Web by content (i.e. Yahoo, AltaVista, OpenText, WebCrawler). Some of these resources are more structured than others; Yahoo, for instance, allows users to limit searches by category, or to browse through indexed sets of Web sites. More commonly, searchers are presented with a dialog box similiar to this:

What rules apply to the query a user enters into such a box? Is boolean searching allowed, and in what form? Since every search engine has limitations on what part of the web is indexed, what are the limitations of the underlying database? What effort has been made to insure that sites indexed are current and appropriate? In many search engines, this information is either unstated or buried deep within the site.

Occasionally, a search engine will offer a more sophisticated search form, like the one below, which is modeled after the OpenText power search form:




Observation (Wallace and Kupperman, 1996) has indicated that students rarely take advantage of the features offered by these more complex search forms and that when they do, their uses of advanced querying capabilities does not result in effective search techniques (Greene, Devlin, Cannata & Gomez, 1990). Having students engage in unstructured searching on the Web may be unproductive (Wallace and Kupperman, 1996). My own observation has been that students involved in unstructured searching spend far more time trying to use the search engines than they do interacting with content pages.


This Web site and presentation copyright 1997, The Regents of the University of Michigan. Site created by Jon Margerum-Leys. Duplication by permission only, please.