I think the reason no one answered is that there are so many different factors that could affect accuracy. 6040Z is more like a concept than a spec--a lot depends on how it was put together--though my guess would be that twist between the z-axis and the x-axis is going to be a major source of its error, depending on what you are cutting. As a first pass, one way to resolve that Gordian knot would be to run some benchmarks. Measuring the results would quickly get you some answers about the strengths and weaknesses of your particular machine--al beit on those particular benchmarks--and thereby narrow down the focus as to what things to focus on. So, I propose that as the "better way."