> I think that the author implicitly relies on the unstated assumption that for business related databases the vast majority of non-integer data is money That

> I think that the author implicitly relies on the unstated assumption that for business related databases the vast majority of non-integer data is money That would certainly result in a much shorter article. "Hey, databases deal with money all the time. They should have a fixedpoint integer type for money values. Floats are a bad choice". If thats all they wanted to say, I wish they just said that.