Text search engine

One of the challenges with textual data is the ability to perform fast contextual searches. Editors, such as vi, have the ability to quickly pattern match HUGE amounts of text. Grep is also prolific in the amount of text it can parse. I am looking for a methodology to tie to the front end of a database which will augment the text search, or MATCHES, capability of a database engine. Particularly, I would like to add the functionality of finding two words which are at most N words apart. (Words are close to or near each other). This is the base requirenment of the add on functionality, but I don't want to specifically restrict myself to this. What is the 'current' state of technology(software) in the field regarding text searches / pattern matches? Related functionality includes, but is not restricted to: - sounds like (SOUNDEX type search) - means the same thing as - case sensitive/insensitive - skipping hyphenation/punctuation of raw data - most common words used in text Steve Wallick Harris Publishing Systems Corp spw@sur1a.ess.harris.com