English 200 Tips on Search Engines

Don't Search Far and Wide:

Don't use a single word in a search query or the search will turn up everything under the sun-most of it useless. Search for in-context phrases for more on-target results. Usually, you can search for an exact match to a text string when you enclose it with quotation marks. If the engine supports it, combine multiple phrases with the Boolean AND to further refine your query. Check your engine's instructions for the most accurate details.

Stay Away from Spamdex Time-Wasters:

When framing your query, avoid common "spamdex" terms-words that Web site marketers often add into hidden parts of a page's HTML coding. This is done so that these pages wind up on the top of query results lists and get numerous hits. Spamdex terms include sex, free, shareware, Web and Windows.

Stop Stopwords :

Search engines routinely exclude stopwords, terms that are very common on Web sites. Query Alta Vista for the word "computer," for instance, and you'll be told that the word was not found. Most engines also ignore one- or two-letter words, and words beginning with a number. Some engines return a list of terms they ignored in your query. Make a note of them to save yourself time in later queries.


Unusual words are more likely to return valid search results than common words. If you must search for a single word, make it the most unusual word possible. Otherwise, you'll spend a lot of time traversing unrelated sites.

Search Engine Syntax:

It is important to note that not all search engines follow the same syntax. For instance, Alta Vista treats John Keats as two separate words with a forced exact case match, while Infoseek treats John Keats as a single name or title. Unless you are dealing with names it is best to use lowercase.

These switches will work with most search engines:

"john keats" will find documents with this exact phrase without regard to case.

+python +monty -snake will find documents containing the words python and monty but not those containing snake.

Cyber* is a wildcard and will find documents like cybernetics, cyberspace, or cyber-anything.


Created by Bruce Matsunaga

Updated: April 30, 1997