Impetus

I seem to spend a lot of time reconciling tables to find differences, like ensuring a list of contacts (mapping names to phone numbers) is up to date. This is not in itself difficult, provided elements that refer to the same thing are the same string, but this is not always (or usually) the case. Variations in spelling, abbreviation, etc. make it difficult for a computer program (using simple string matching) to tell if a given pair of entries is different because (e.g.) one is outdated or they're spelled differently.

tablediff is an attempt to mitigate this problem by applying clever heuristics.