[letney]

> "you can't even represent 0.1 in IEEE floating point"

Not true, the IEEE 754-2008 standard provides a Decimal Floating Point specification in binary.

http://en.wikipedia.org/wiki/Decimal_floating_point#IEEE_754...

[kps]

Ah, well, you're the best kind of correct. IBM even have instructions for it.

I'll counter with the Ferranti Orion (1959) which had mixed-base operations (primarily for £sd accounting).