Home

Published in Link-Up Magazine

A Short Primer in Conducting Searches


by Wayne English

Everything and anything you could want is on the Web.

That is both its strong point and its weakness.

How do you find what you want in all that data? By using a search engine and a targeted search.

Come on, 'fess up. How many searches have you made that returned hundreds or even tens of thousands of sites? After you read this article, that will never happen again. From now on the Web will be your servant. You will not have to deal with a mountain of data to get the kernel you are looking for.

Constructing the search

You will use the following characters and Boolean operators to construct a search. Not all of these characters and operators are supported by all of the search engines. There are far too many search engines on the Web to make such a sweeping statement. You will find information about the most popular search engines and the search operators they employ in the next section.

As there are many ways to phrase any search, always try to always use words that are as specific as possible to what you are seeking. If your search is unsuccessful, do not give up. Keep trying different search phrases and search engines. Here are the tools you will use:

+ The plus sign. For example: Cambridge+hotels. Here Cambridge is capitalized because it is the name of a town and we are sure that it will be capitalized on the Web.

- The minus sign. For example: cars-Ford.

" " Quote marks. For example: "fuel cells." This finds the string "fuel cells," i.e. the word fuel, the space character, and the word cells.

() Parentheses. Used to group portions of Boolean searches. For example: colleges AND (Connecticut OR Vermont).

AND The Boolean AND. Example: gasoline AND kerosene.

OR The Boolean OR operator. Example: cars OR ford. This search will find all sites with the words cars or the word ford.

NOT The Boolean operator NOT finds sites that include the first word but not the second. For example: cars NOT ford.

AND NOT Same as NOT.

The pipe command. This will filter your search returning only what passes through the filter. For example: cars | ford. Of all the sites containing the word cars you will only see those that contain the word ford.

NEAR Example: ibm NEAR/25 sun. This search will return only the sites where "ibm" and "sun" appear within 25 words of each other.

ADJ Example: ibm ADJ sun. ADJ finds only those sites where "ibm" and "sun" are next to each other in that order.

site: This command will search for the Universal Resource Locator (URL) of a specific site and its related pages. For example: site:cnn.

title: This command searches for a specific title. For example: title:Cable News Network.

Fuzzy search capability. Choose this whenever you can. Fuzzy technology makes sense out of, well, fuzzy input.

Capitalize the Boolean operators, but leave the rest of the search words in lower case unless you are sure the data you seek is capitalized (see the Cambridge example, above).

Search engines

A search engine is software that queries a database. No single search engine scours the entire Web for information. Not yet anyway. Your search will be submitted against one, or more in the case of multi-threaded search engines, databases. That's why it is important to use various search engines if one or more does not find what you want. A search failure does not mean the data is not out there. It just means the database queried does not contain it. Open any of the search engines listed here and search on: search engines. You will find a wealth of tools just ready and waiting to go to work for you.

Your search needs are simple: Find the data and return it. What could be more basic? The problem comes when your search brings home 17,000 possible sites. This is not helpful. In fact more than 20 is not good. You can easily reduce your hits to 15 or 20 by conducting a structured, targeted, specific search. The tools listed above will get the job done for you. For example, you might search for: "search engines" + "online catalogs." Perhaps you are looking for special tools, office supplies, or special protective gear. The search is free and you will find suppliers from across the country and around the world.

Multi-search engines

These engines will search multiple search engines from a single search request.

MetaCrawler

(http://www.metacrawler.com) is perhaps the most sophisticated but at a cost of increased search time. It supports quotation marks, +, -, AND, NOT, NEAR, ADJ, OR, | (the pipe command), site, and title. MetaCrawler searches AltaVista, Excite, Infoseek, Lycos, Webcrawler, and Yahoo!. You'll be able to shop online, get stock information, and conduct research on companies.

SavvySearch

Located at http://www.cs.colostate.edu/~dreiling/smartform.html, the SavvySearch search engine uses keyword searches. Typically this search engine ignores small words such as it, the, for, etc. Numbers and most symbols are allowed. There is no need for AND and OR as these operators are inserted for you depending on your query selections. Initial search criteria can be focused on WWW Resources, Software, People, Reference, Commercial, Academic, Technical Reports, Images, News, or Entertainment. SavvySearch can display in a couple dozen languages. Web indexes include AltaVista, Excite, Infoseek, Inktomi, Lycos, OpenText, Webcrawler, and HotBot.

General search engines

AltaVista (http://altavista.digital.com/) is operated by Digital Equipment Corporation. AltaVista claims the largest Web index with eight billion words. AltaVista supports a powerful array of search tools, including AND, OR, NOT, NEAR, date ranges, ranking of results, quotation marks for exact phrases, and more. See their excellent help page at http://altavista.digital.com/av/content/help_advanced.htm.

Excite (http://www.excite.com/) features Web and Usenet searches and Web site reviews. The Excite opening page offers categories from which you select to begin your inquiry. Also there are a people finder, stock quotes, maps, sports, and horoscopes. You can also book flights here. For a power search, simply click on the Power Search link available on the opening page.

Lycos (http://www.lycos.com/) is rated number one in relevancy and number of hits by PC World. Supports +, -, and quotes for phrase searches. Here you can access the Web, UPS tracking, stocks, recipes, pictures, sounds, books, and personal home pages.

Infoseek (http://guide.infoseek.com/) is a broad service with the Web, Newsgroups, and Usenet. Search tools supported are quotation marks, +, -, | (the pipe command), link, and site. Infoseek offers numerous categories such as Automotive, Business, Careers, Computers, Personal Finance and others that you can use to zero in on the information you want.

Open Text (http://index.opentext.net/) indexes every word of every Web page the Open Text has visited—some 21 billion words and phrases. Their advanced search engine does all the work for you. Visit their power search page at http://index.opentext.net/main/powersearch.html.

WebCrawler (http://www.webcrawler.com/) was started at the University of Washington and is now owned by America Online. Its opening page offers services and selectable categories that you can click on. Supported search tools are AND, OR, NOT, NEAR, ADJ, quotation marks, and parenthesis.

Yahoo! (http://www.yahoo.com/) is considered the granddaddy of the field. Regarded by many as the best and most complete Web directory, Yahoo! provides Yellow Pages, Pele Search, Maps, Classifieds, Personals, e-mail, and more. You can also access World Yahoo for other country information and Yahoo! Metros for data from specific cities.

Search.com (http://www.search.com) provides search forms and links in 20 different categories: Arts, Automotive, Business, Computers, Directories, Education, Employment, Entertainment, Finance, Government, Health, Legal, Lifestyle, News, Reference, Science, Sports, Travel, Usenet, and Web.

To access additional search engines covering a wide range of data services, surf over to http://cuiwww.unige.ch/metaindex.html. You will find a list of search engines from the University of Geneva.

You will see that by using these techniques, the Web will change from a confusing flood of data to a premier, targetable, and amazing search tool.