Fro CERTAIN types of string info, there is technique "Soundex", which does a good job of locating 'near' matches to single wiords and short phrases. It was developed by the U.S. Gov. way long ago to sort names on the "Sound Index" of a name. There is a lot of info on the technique available, although most of it is rather old. The process basically reduces the 'english' language to an index on the SOUNDS in the Word(s), so that "Mac" And "Mc" and "Mak" And ... Are all grouped as the same Index value. To use the technique effectively, you need to add a soundex field to the recordsets where you would use the soundex for a look up, and generate the soundex for each record. To look up the name, you would enter the "name" and create the soundex for it and then look up the soundex - but return the actual name(s). It works best when yuu return the list of names which are "near" the soundex, I ususlly use a combobox or list box and have the user pick from this list. If you are just looking up english language names or 'phrases' of three or less words, many of the routines found in the public domain are adequate.
If you need to look up more complex strings, you may need to either modify the standard soundex model or re-think the db structure. The original routines were designed to accrpt ONLY names, without titles, salutations, abreviations, certifications ..., so if you have a "Whole Name" field, the titles (Mr., Mrs., Ms. Dr. ... ) will have a major (NEGATIVE) impact on the soundex indexing, so you would either need to Modify the soundex calculation to remove this 'clutter', or re-design the db to seperate the name field into sub-sets, like title, first name, middle name (Inital), Last name ...
MichaelRed
mred@att.net
There is never time to do it right but there is always time to do it over