Stumbled accross this when looking for a way to calculate a thing in stats; it's an orthographic similarity equater.
For example, type in tobar and then tobair and see how similar they are (out of 1000 points).
I don't know if any practical use can be found on it's own (with other programs it could be used to test the 'phoneticness' of a text based on an extra set of rules).
Anyway, enjoy!
http://users.ugent.be/~rhartsui/Applet1.html