Comment 2 for bug 259221

Could you not just parse the diffs and extract words separated by punctuation and white space? This would be independant of whatever language the source file was written in. Obviously one would want to filter common words like and if not or etc but otherwise I don't see why this can't be done generically.