Magazine article Computers in Libraries

Online Treasures

Magazine article Computers in Libraries

Online Treasures

Article excerpt

Janet L. Balas is library information systems specialist at Monroeville (Pennsylvania) Public Library. She can be reached by email or

Exploring Some New Search Tools for Librarians

As I am sure you can guess, I spend a significant amount of time online. Sometimes I'm just poking around following up on some of my more frivolous interests, sometimes I'm looking for ideas for upcoming columns, and sometimes I'm searching for something very specific either for myself or for a reference question at my library. I find the Web to be the most frustrating when I'm searching for a specific piece of information. Of course, there are the search engines, but often they give either too few or too many results. Like everything else on the Web, technologies for searching are constantly changing and adding new features, and this month we'll take a look at some new technologies that have either just made their debut or are still under development.

A New Technology to Search the Web Clever-ly

A search technology that may hold great promise, but is not yet ready for release, is the Clever project from the IBM Almaden Research Center. When using the familiar search engines, users are frequently overwhelmed by too many results, most of which are only marginally relevant to the topic. The Clever project seeks to alleviate this problem by returning only the most authoritative information on the subject.

Clever extends the search algorithm developed by Jon Kleinberg. This algorithm finds authoritative sources on the Web (referred to as authorities) and also sites (referred to as hubs) that are compilations of authoritative sources. His Hypertext-Induced Topic Search algorithm, referred to as HITS, works by first using a text search engine to find pages matching the requested subject, and then adding all the pages that either link to these pages or are linked to from these pages. After building this pool, then only the links between the pages are used to determine the best authorities and hubs. By using only the links, the algorithm uses the judgment of individuals who have created these links. Clever extends this algorithm by also using the content of the Web pages to determine which sites are relevant to the query.

In addition to the explanation of the Clever Project, the Web site has links to articles that describe the ongoing work of the project, links to some sample searches, and links to coverage of the project on other Web sites. One very interesting link on this page is to an article from The Industry Standard (available on the Web and in print) entitled "The Future of Search Engines," which discusses Clever and two other new search technologies, DirectRit (See my previous column, "It's the Little Things That Count," Computers in Libraries, December, 1998, pp. 35-39.) and Google, another new search technology worth a look.

Exactly What Is a Google?

The answer to the above question can be found at the Google Web site. Even though Google is considered a technology still under development, a beta version is available online. Google technology, which was originally developed in Stanford University's Computer Science Department, bases its algorithm on the links between pages. If pages deemed "important" point to a site, then that site is also considered "important." This ranking, combined with sophisticated text-matching techniques, is used to determine a set of quality results to a query.

Google developers are so confident of its ability to find the best answer to a simple query that an "I'm feeling lucky" button is available to take searchers directly to the one best result of their search without having to wait for a longer list of results to load. Of course, if a searcher doesn't feel that confident or lucky, a traditional results list can be selected instead. Visitors to the Google site can learn more about the technology and pick up some Google search tips by clicking on the More about Google link. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.