@matrix There is a use; real-world measurements where there is noise in data, inaccuracy in sampling technology and much more.
And those values are rarely used in Internet and Enterprise apps, so I agree "why is it there".
The abomination is programmers using floats as if it is a precise value. Instead of seeing immediate rounding problems in fixed-point arithmetic, we get obscure bugs from aggregated errors.
It is human to err, but to really fuck up, one needs a computer.