Skip to Content

Search Engines

Printer-friendly versionPrinter-friendly versionSend to friendSend to friend

Tracey Stanley writes about "Ask Jeeves", a search engine which processes natural language enquiries.

Search engines are getting more sophisticated all the time. It's likely that we'll soon be seeing the rapid emergence of more 'intelligent' search tools which offer features for personalisation and tailoring, more effective searching synatax and more effective methods of tracking down information on large databases. Such tools are already starting to emerge in the search engine market in response to users' needs for more sophisticated tools to help them make sense of the growing morass of information on the Web.

One such tool is a meta-search engine called Ask Jeeves. Ask Jeeves[1] has attempted to cross the divide between large automated indexes such as Alta Vista and Excite, and the smaller hand-crafted services such as Yahoo! by offering a meta-search tool which utilises human intelligence in the filtering and organising of resources. Search Ask Jeeves on a particular topic, and you'll get a list of search results. However, what you'll also be presented with is a question. For example, if you search Ask Jeeves for information your favourite novellist it will come up with a large set of results, as would usually be expected from a meta-search engine. However, what Ask Jeeves will then do is preface these results with a number of questions; for example, "where can I find a review of the latest Iain Banks novel?" or "where can I get hold of a copy of The Crow Road?". Selecting one of these questions will take the user to a collection of web sites that seek to answer the question - perhaps an online bookstore or a literary electronic journal - thus helping a user refine their search effectively without too much hard work or effort.

These questions, and their answers, have been manually selected by human editors who scan resources on the Web on a daily basis to build up a knowledge base of information about sites which might be used to answer common questions. The questions and the web pages which answer them are then stored as a series of templates in the Ask Jeeves knowledge base, and keywords and concepts in a search string are matched against them in order to retrieve the questions, and their corresponding web sites. Ask Jeeves represents a clever model of an application using knowledge management techniques in order to better organise disparate information sources. It draws upon the expertise of experienced human web searchers, and encapsulates this expertise in a database so that it can be put to use by others.

Ask Jeeves uses Natural Language Processing for searching, which means that questions or search keywords can be typed in plain english (for example, "where can I find reviews of the latest films" rather than +"film reviews" +latest - which might be required by the more traditional search engines). This should make it easier and more intuitive to use than many other search engines.

Ask Jeeves claims that, in contrast to the hundreds of thousands of irrelevant sites retrieved by most web search engines, users of their service will be directed to at most two or three relevant sites[2], thus cutting down dramatically on the time spent trying to track down relevant information sources.

Ask Jeeves can also interact with a user in order to better define a question that is too general or vague to provide an immediate answer from the knowledge base database. For example, a question such as "what is the email address of Bob Brown" might result in a question asking where Bob Brown lives which would help to narrow down the results.

References

[1]Ask Jeeves http://www.askjeeves.com

[2]Ask Jeeves, Help Pages http://www.askjeeves.com/docs/HelpFrame.html

Author Details

Tracey Stanley
Networked Information Officer.
Library, Computing and Media Services,
University of Leeds
UK Telephone +44 113 233 5569
Email: T.S.Stanley@leeds.ac.uk
Date published: 
19 September 1998

This article has been published under copyright; please see our access terms and copyright guidance regarding use of content from this article. See also our explanations of how to cite Ariadne articles for examples of bibliographic format.

How to cite this article

Tracey Stanley. "Search Engines". September 1998, Ariadne Issue 17 http://www.ariadne.ac.uk/issue17/search-engines/


article | by Dr. Radut