Published May 23, 2001
Beginner or novice researchers’ first universal experience is Web searches don’t work. This Web stuff must be a bunch of bunk; thousands of irrelevant and useless matches. The tool is not broken; the user has mistakenly griped the hammer from the wrong end.
As of July 1999, an estimated 800 million Web pages were publicly indexable. One year later, it had doubled to 1,580 million, according to NEC Research Institute. With all this content, finding relevant information on a chosen topic shouldn’t be that difficult, should it?
The Web, a small portion of the entire Internet, is not entirely cataloged nor can it ever be—the fastest computers in the world are straining to stay current with the automated task. No more than 17% of the Internet’s Web pages are listed in any one of the 50-or-so major search engines (SEs). All toll, 2000 search engines may exist worldwide, most relating to very narrow topics.
Some prefer the Web surfing method of researching. Web surfing is the ineffective practice of randomly clicking upon subsequent Web page links leading to who knows where. However if time and patience were sufficiently abundant, a Web surfer clicking upon one link every minute for 8 hours a day, every day for five years, will have only seen one half of one tenth of 1% of all there is. This analogy is ludicrous, of course, because the size of the Web in that five years will have increased 32-times (doubling every year).
For better or worse, search engines are the principal repositories for searching and finding Web pages matching given key words. Each SE acquires and organizes their lists uniquely. Individuals are allowed to manually submit the page information or robots, called spiders, roam the Web collecting information of Web pages links to other links endlessly.
Some search engines organize their vast collections into topics categories. Others, called meta engines, organize queries into a common list from indexes not their own. All will extract matches from databases for your typed-in keywords.
Searching any site for the words containing Livingston and Montana might rank a response list that is either, not both. The relevant list will improve if the most important word is typed-in first: Montana then Livingston. Otherwise, their dumb computer will start with the first word and return links to Livingston from New Jersey, Louisiana, Texas, or any of the five other states or six countries with the same city name. You may see the most obscure list of links relating to the person, a rock band, or someone’s pet named Livingston.
The current evolution of search engines have adopted a radio-button of three basic styles: any words, all words, exact phrase. Choosing the any words will list all indexed pages on that SE that contain any of the search words and yield an immense list of results. The all words lists only pages that contain every keyword. The exact phrase style, like a person or place, will return only pages with that sequence of exact words, ignoring all others.
What about a combination of these to narrow the search for a shorter more meaningful list? You bet. Look for a link from the SE’s start page that’s labeled Power or Advanced search. Read the choices carefully and proceed accordingly. Use lower case except to capitalize proper names; avoid all caps. Omit articles such as: the, of, to, and, for. The use of commas, quotes and/or parenthesis can noticeably help your results, or may be ignored dependent upon the whims of each SE. Read the Tips link to learn about the styles your favorite SE employs, and practice using them.
How about searching without a SE? You sure can. Try this: from your Web browser’s address (URL) field, type a question mark (?)—sans the parenthesis—a space, your keyword, and then the return on the keyboard. For multiple word searches in Netscape, omit the question mark; or in Internet Explorer, replace the question mark with the word go then your keywords and the return key. Slick huh?
Check back next week for part 2 of Search-Engine Tips.