Duane,

You are right. You take the biggest deviation by absolute value, double it and this is your actual (similar to TPD). What they were telling you makes no sense. If you go by their logic error of 0.0mm should be reported as 0.5mm.

Ray is correct that it is more useful to report +/- deviation, but this is not what the call out asks for. And I'm completely lost at Geoff's article.

George K.
Sitius