REAL datatype math problem

[CanaanP 1]

Hello, I am working on a project where I measure material with accuracy to the tenth of a mm. I also store a value of blade thickness that will cut that material that is input as mm.m If I have a piece of material that is 119.2, my blade thickness is 3.2, then I should wind up with 2 58mm pieces. However, after subtracting one value from another I wind up with 57.999996 as the value of my second piece. I had asked for advice from another source on how to handle this and they said to use an implied decimal. Meaning, use DINT as datatype and just add a zero to the end of each value, then have HMI put the decimal there. Is this really the best solution? Or is there another route I can take where my datatypes are the proper type and I can limit accuracy to a tenth?

[b_carlton 61]

Of course the problem is that floating point can't represent the 3.2 exactly. Though it can represent the 119.2 exactly. Subtracting you get an inexact number close to 116.0 but not exactly. As long as your single decimal place is exact then using whole integers (multiplying the original number by 10) may be the best route. The HMI can supply the proper positioning of the decimal point.