The Internet Encyclopedia (Volume 3)

P1: IML

Wisman WL040/Bidgoli-Vol III-Ch-59 August 14, 2003 18:3 Char Count= 0

726 WEBSEARCHFUNDAMENTALS

Lists

Catalogs, directories, and lists are human-organized sub-

ject indexes of Web information. A list arranges subjects

into a hierarchy that allow exploration from a general

topic such as “education” down to the more specific

subtopic of “graduation poems.” On most lists, humans in-

dex the subjects providing control and intelligence to the

subject hierarchy organization, arguably yielding less but

higher quality information than current automated cata-

loging systems. The hierarchical approach works well for

exploratory search when one is familiar with the search

subject and has a good sense how the subject fits within

the hierarchies of a larger subject.

Finding a subject in a hierarchical list can be diffi-

cult when the searcher’s idea of subject organization dif-

fers from that of the person creating the list. For those

cases, a keyword query can search the subject list as a

way to pierce an opaque subject hierarchy; a query such

as “graduation poems” will generally locate the same in-

formation as manually moving through the “education”

subject hierarchy. If the query fails, most list services au-

tomatically send the query to a Web search engine; for

example, the Yahoo list currently collaborates with the

Google search engine for Web searches of failed subject

queries.

Search Engine

A search engine matches query words with words on Web

pages and lists the pages containing the matching words.

Entering the query word “zucchini” returns a list of pages

containing the word “zucchini.” In practice, Web search

engines examine vast numbers of pages while automati-

cally calculating page importance to rank each page from

highest to lowest to reduce quantity and improve infor-

mation quality. Accurately determining page rank based

on a few query words is challenging and sometimes pro-

duces completely irrelevant results. Entering the single

word “zucchini” will find any page mentioning “zucchini,”

from gardening to cooking to diet; which “zucchini” page

is important is ultimately in the mind of the searcher.

Determining how to rank one document against thou-

sands of others is a key point of competition between

search engines and often the identical search on any two

will produce different results, searchers may need to con-

sult several to have confidence in the results. Metasearch

engines automate searching on multiple sites by send-

ing the search query to a number of search engines and

creating a fusion of the results. Overall, search engines

are best suited for existence and comprehensive searches

where the quantity of information is of prime impor-

tance.

Portal

A portal attempts to make using the Web easier. Portals

attract a broad range of visitors by serving as informa-

tion department stores with a variety of services, such as

e-mail, forums, search engines, subject lists, and online

shopping malls. Portals are also distinguished from pure

information sites by the degree of personalization possi-

ble for information sources by combining search to pro-

duce personalized services for local weather, stock port-

folio tracking, calendars, and search categories. Portals

often integrate external search engines with internal sub-

ject directories and other information tools to improve

quality and a seamless interface to a set of multiple ser-

vices. Portal tool integration and ease of use has trans-

formed the Web into a household appliance, one with

obvious benefits and easily understood and used by the

broader population (Hock, 2001). Since 1998, portals have

been a major force driving the rapid growth of the Web.

America Online (AOL) was among the first portals and

has one of the most extensive lists of services and degree

of personalization. Today, most major search engine and

list sites now offer many of the same services to attract

and keep revenue-generating visitors. Other portals target

visitors with narrower information and service needs; for

example, a financial service portal would provide a range

of personalized financial services along with focused fi-

nancial information. Even as competition blurs the dif-

ferences between portals, lists, and search engine sites,

information in some form remains the primary lure to

attract visitors to each.

How to Use Search Engines

What is the best search strategy? For a clearly defined

subject area, use a site that maintains subject lists. On

Yahoo, one can find “weight loss” carefully placed in the

“Health > Weight Issues > Weight Loss” subject hierarchy.

The “Weight Loss” category then has seven subcategories

ranging from liposuction to diets and a list of some 30

other Web sites devoted to weight loss. Human thought

in placing each subject in the organization is obvious as

each subject relates in some way to others nearby. In com-

parison, a search engine can find over 1 million pages with

the words “weight loss.” Although the search engine cal-

culates the rank of each page, there is no guarantee that

one page relates to another except through the common

coincidence of the words “weight” and “loss.”

When are search engines the better choice over lists?

With no clear definition of the subject area, subject lists

are of little value. Consider locating genealogy informa-

tion for a family. Unless the family is already famous, find-

ing a subject list devoted to it is doubtful, but entering the

family name on a search engine would likely locate dozens

of individual pages and Web sites. A search engine is better

suited for finding comprehensive information that does

not fit into a clearly defined category.

How to Search

One of the main problems with search engines is they are

much better at finding a high quantity than a high qual-

ity of information. To understand how this affects search

strategies, consider that search engines operate by key-

word searching, searching for pages containing one or

more of the words entered as a query. To research diet

plans, a searcher might enter the query “weight loss” and

the search engine might return several million page ref-

erences, each containing the word “weight” or “loss.” Can

search be limited to pages that are useful? The answer is

yes, but because search engine strategies purposely dif-

fer, there is not a single protocol for defining a search.

There are, however, some common approaches to improv-

ing search success.