Hi,

I work in a company which processes data from the local stock exchange. A problem we are running into is that the exchange places an upper limit on the number of significant figures in prices disseminated (10), but the decimal point can float.

This, obviously, sounds like the perfect job for double. However, we can't afford to lose any precision when it comes to the prices.

Eg. double a = 2.30 will set a = 2.9999...9 not 2.30. That is unacceptable. The problem, of course, is that doubles are stored with a floating point with respect to base 2, not base 10.

Has anyone else come across this problem? My solution was to write a base 10 floating point class (essentially coupling some integral storage variable with an integral variable containing the location of the decimal point), but it's not an especially attractive solution. Still, I can't really think of any other way of doing it. I don't really seeing it as being particularly onerous to extend this class to handle arbitrary bases.

Has anyone else ever found themselves in this position? Would this type of code be suitable for boost, or is it too mission-specific? I guess the line has to be drawn somewhere, and this ain't exactly building a rocket...on the other hand, it is tedious enough that no one really wants to do it. So I don't really know.

Damien