Thanks John, I'll check metaphone out, but SOUNDEX() is a fine function for my encoding needs.

This is a different issue. What I want, in effect, is a distance function that uses common typographical errors as a means for determining the metric.

Writing it myself is an option (one I'm trying hard to be disciplined about). I always like to see if there is a $199 piece of shrink-wrap that will do the job. If it would me more than two hours to replicate, then it's a good deal.

I think I've some old text books with some algorithms in them... might be a fun project to create a little VFP library for the purpose. Probably be a nice implementation using the strategy pattern to allow different algorithms to be swapped in and out. See... now you done it... I'm loosing my self discipline.


-lc