We used this when building an app to determine 'meaningful' discrepencies between prelimary radiology reports (done by Residents) and final reports (done by attending radiologists). It's a fun algorithm to implement that isn't as simple as it seems. I think there's lots of uses for it.