You have to relate your results to a percent of tolerance, as I understand it. That with gage R&R. The MSA manual talks about looking at it all as a 'system'. If I understand it correctly, they want to make sure you are considering all the system elements. You may get away with not doing linearity or bias calculations if you can argue the precision of the measurement device far exceeds the tolerance (extreme case --> using a cmm to measure a 0.9 +/-0.1 meter rod). But then, you always know the linearity and bias and stability and uncertainty of the cmm simply because the manufacturer provides them for their products due to their very nature.