Is there such a thing as too much precision?

Yes, I believe it's called 'false precision': http://en.wikipedia.org/wiki/False_precision

You didn't read the article (which incidentally isn't worth reading).

Basically the final point after lots of text is that you want all the precision because sometimes you want to use a number that can be represented exactly in binary float.

Doesn't really address the point that it still makes no difference to the final value.

I think there is a trade off between precision and maintainability. If you can code it and never touch it, I would go for precision. But if you are going to revisit the code often enough I would go for maintainability.

Suggestion to the author: learn about fuzzy logic and than start to think.