Help Formulating a Search
Search Styles

Statistical Searches

Statistical searches are what make this search facility unique. Many search tools search for word matches and record ids but few do this in combination with a search for similar context. When you have an idea or concept that needs to be matched, word searches alone are often of little use. Thus the Statisitcal search is given significant attention.

Searching for similar text returns items similar in meaning to the description given in the query, even when that description is imprecise, misspelled or incomplete.

Word Stems

These searches enable you to find documents related to your search queue even when you do not specify the exact word because the stems of the words are used. A query with the word "mounting" would also return documents with the words "mount","mounted","mounts" and so on. In this search style we include a soundex algorithm that returns records with misspelled words or words that sound like query words. For example, the query "Smith" returns records including "Smithe","Smyth","Smiiith", and "Smth".

Exact Words

Exact word searches retrieve documents that contain a set of user specified words. This search style is quite useful for those occasions when you wish to narrow your search to a few very unique words. A maximum of eight words may be specified. Word searches are case insensitive.

For exact searching, use precise words. If an exact word query returns a very large hitlist, replace the word in the query that is the least specific. Often, replacing the first word will have the greatest effect because we humans tend to specify broad category words first and use later words to subdivide the initial broad word.

Help Formulating Search Queries

The following search capabilities are enabled for this World Wide Web server. Other capabilities may be enabled at a later date.

Weighted capabilities are built into Statistical searches. Documents of relative importance are weighted higher on the hitlist than others.

Root Expansion capabilities are supported for Statistical and Word Stem searches.

Boolean capabilities are supported for Word Stem and Exact Word searches. Boolean allows queries to be qualified using operators that represent AND, OR, and NOT. Nested parentheses are also supported. For boolean operations use:

"&" or a space = AND
"|" pipe symbol = OR
"~" tilde symbol = NOT

If no boolean operators are used, AND is assumed. Without parentheses, words associate left to right. Words inside parentheses are considered first as themselves, then as the group inside the parentheses. Example:

Rush | (April ~4/17/98)

An Exact query like the one above would retrieve all documents containing "Rush" and also all documents containing "April" but not "4/17/99".

Bridge characters will usually be interpreted. Characters such as . / _ - will usually be considered in the search query resulting in finds like "my.doc" "/u/home/me" "yourfile_ps" "built-in".

Case sensitivity is not supported. All searches are case insensitive.

Word order is not significant.

Trivial words are ignored. Examples include "and or the also this their" and many more.