Evaluation of Selected Internet
Search Tools
Search tools or search engines attempt to retrieve resources (usually Web pages)
containing the terms that describe your topic. Although not all search tools permit the
use of the Boolean operators AND, OR, and NOT, most do use relevancy ranking for retrieval
and display. The documentation accompanying many of the prominent search tools is replete
with claims about speed, database size, search features, accuracy, and precision. The
truth is, even the smaller search tools will sometimes retrieve useful material that the
bigger ones miss. There is also no unanimous agreement on how to best calculate the size
of a search tool's database. Some tools only count the Web pages they have actually
retrieved and indexed. Others also include the number of links mentioned in those
retrieved pages. Try several to ensure a more complete search. Almost all the search tools
listed below provide helpful search tips. Use this information to take full advantage of
each search tool! You may also wish to examine the Search Engine Watch site for more
valuable information and resources on search engines. There is also a guide to Getting the
most out of Guides and Search Engines that may prove useful in your work or research.
- AltaVista
- http://altavista.digital.com/
AltaVista says it provides access to "31 million pages found on 476,000 servers, and
four million articles from 14,000 Usenet news groups."
Databases: WWW pages, Usenet news groups
Search features: Full text; Boolean operators AND, OR, AND NOT; restriction of
search to certain portions of the web page; phrase searching; case sensitive;
truncation/wildcard; use of parentheses to group search expressions.
Results: Results can be viewed in standard, compact, and detailed form. Rankings
criteria must be specified for "Advanced Query" mode.
Comments: AltaVista's ability to target those WWW pages that link to your site is
one of its most popular features. Extensive search tips for features not mentioned above
are available.
- DejaNews
- http://www.dejanews.com/
Databases: Archived Usenet news groups "ranging back to March, 1995."
Search features: Select "Power Search" to take advantage of the full
range of search options including the Boolean operators AND, OR, AND NOT, and NEAR. Other
features are phrase, proximity and field searching.
Results:
Comments: Updated daily. An extensive help index is available.
- Excite
- http://www.excite.com/
Excite's producers say that their NetSearch database is not padded by including the number
of links embedded in indexed pages.
Databases: Two weeks of Usenet news group postings, Usenet classified
advertisements, and NetReviews, a collection of over 60,000 brief Web site reviews.
Search features: Full text of WWW pages, excluding common stopwords; use of Boolean
AND, OR, AND NOT; parentheses to group portions of Boolean queries together; "query
by example" option finds pages similar to the search result you like best.
"Concept-based search" (not available with above search features) attempts to
retrieve documents with your exact query words and related words.
Results: The annotated results are returned in order of decreasing relevance.
Comments: Select "Advanced Query Language" for tips on how to fine-tune
your search. The index is rebuilt approximately once a week.
- Galaxy Search
- http://galaxy.tradewave.com/search.html
Databases: WWW pages, Gopher titles, and Telnet resources from the Hytelnet
hypertext telnet database.
Search features: Full, title, and text searches, retrieval of ANY or ALL search
terms, automatic truncation.
Results: Output length (annotation of results) may be selected.
Comments: Use the name of the sub-directory included with each retrieved item to go
seek other similar items in Galaxy. Provides some search tips.
- HotBot
- http://www.hotbot.com/
HotBot is powered by the Inktomi search engine.
Databases: WWW pages, Usenet news groups.
Search features: Retrieval of all the words, any of the words, phrase, person, and
URL. For extra options, select the "modify" and "expert" buttons.
Results: The annotated results are returned in order of decreasing relevance.
Comments: Excludes common stopwords.
-
- InfoSeek Guide
- http://guide.infoseek.com/
The InfoSeek Guide database counts unique Web pages, not the URLs (or links) mentioned in
those pages.
Databases: WWW pages, Usenet news groups, InfoSeek Select Sites (reviewed),
directory of companies, e-mail addresses, one month of news stories, FAQs
Search features: Full-text, excluding common stopwords; case sensitive; symbol and
numbers recognition; phrase searching; proximity; use of the + (plus) and - (minus)
symbols to designate required and unwanted terms.
Results: The annotated results are returned in order of decreasing relevance. Some
results include a list of related topics and news groups.
Comments: InfoSeek endeavors to be selective in the pages it indexes so as to
reduce the number of duplicate and irrelevant pages retrieved. For best results, use the
search tips.
- Lycos
- http://www.lycos.com/
Lycos merges the results of its continuous WWW sampling into its catalog on a weekly
basis. Its database includes the URLs and links (some Gopher and FTP) just mentioned in
the Web pages Lycos has actually indexed.
Databases: WWW pages, some FTP and Gopher links, A2Z (the most frequently visited
sites arranged by subject) and WWW reviews by Point Communications
Search features: Ignores common stopwords; searches the title, headings, links,
keywords, and first 20 lines of Web pages; default search assumes OR between search terms;
search form available via the "Custom search" option allows you to choose match
strength (loose to strong), a match of ALL or ANY terms, and three levels of display
detail. Also allows narrowing the search to sounds or images.
Results: Results are listed in order of decreasing relevance.
Comments: Provides extensive search tips.
- The Open Text Index
- http://index.opentext.net/
The Open Text Index indexes the full text (even common stopwords) of Web pages.
Databases: WWW pages, Usenet news groups (Deja News)
Search features: Choice between Simple Search and Power Search; Boolean AND, OR,
BUT NOT, NEAR, FOLLOWED BY; limit of search to certain parts of Web pages such as URLs,
titles, headings.
Results: Special features allow users to view their search terms in context and
find pages similar to the ones they like best.
Comments: Open Text does not truncate pluralized words. Be sure to use the search
tips.
- SavvySearch
- http://www.cs.colostate.edu/~dreiling/smartform.html
SavvySearch is a multi-threaded tool, designed to take your query to several search tools
at once, gather the information, and return the results.
Databases:WWW pages
Search features: Using a common interface for several search tools prevents the use
of certain special search features available at some sites.
Results: Option to select length of format. Results are arranged by name of search
tool. Presence of annotation varies depending on the original search tool
Comments: Due to its popularity and the problems inherent in contacting multiple
search tools, SavvySearch can be difficult to reach during the day.
- WebCrawler
- http://webcrawler.com/
WebCrawler allows document title and content searches of its submission and
robot-constructed database. The database consists of both explored and unexplored Web
pages.
Databases: WWW pages
Search features: Words automatically stripped of their endings; use of Boolean AND,
OR, NOT; phrase searching; adjacency; parentheses for grouping searches; proximity
searches.
Results: Search results are returned in order of decreasing relevance in an
unannotated, easy-to-browse list.
Comments: Useful search tips and examples available.
- Yahoo! Search
- http://search.yahoo.com/bin/search/options
Databases: WWW pages, e-mail addresses, Usenet news groups.
Search features: Matches can contain all or at least one of the search terms,
select substrings or whole words, select the number of matches displayed per page. Yahoo!
is not case sensitive.
Results: Yahoo! returns a list of categories that match your keywords, end-sites
that match your keywords, and names of categories where those end-sites are listed.
Comments: Yahoo! searchers may use the name of the categories included in search
results to go seek other similar items.