Non-alphanumeric characters

Thanks for this tool - just hitting minor problems with non-alphanumeric character for matching purposes. For instance...

Say the term is "Harry". Extractor fins a match where the body includes "Harry.", "Harry," and "Harry:" but not "Harry!" or "Harry;".

Would it not be better to run a NOT match against "a-z, A-Z, 0-9" instead? This would define lots of other characters such as ampersands, plus signs, underscores and apostrophes as non-alphanumeric tags. Would be more consistent and easier to understand than current version.