@martin-2 Bastable: Thanks for your kind words, and thanks to those who proofread various drafts of my article, including Kellie Sanchez, Dan Hess, Eric Freeman, Ernest Cook, and Chris Anderson! (Sorry I didn't thank y'all in the article itself.)

I have also implemented systems such as you describe, that pre-process the data and combine it in various ways, to make "fuzzy" searching quicker and more flexible. But for this article, I wanted to present a method that can be dropped into any database with a minimum of work, and still have acceptable performance when "fuzzy" searching on one field.