Excuse the irony, but there seems to be more than one interpretation of what "shall not be subject to more than one interpretation" means. Dave's statements (and I'm not sure that I understand them correctly) indicate that it means that a given feature could be toleranced in more than one way by the designer.
To me, dimensions and tolerances should not be subject to more than one interpretation when applied to a real (i.e. imperfect) part. So that each dimension, and more importantly, each tolerance, can be defined on a real part in a unique and unambiguous way. Measurement uncertainty will always exist, but that is a separate issue. We're talking about specification uncertainty here, where the specification itself is open to multiple interpretations.
In order to truly obey the statements in Y14.5 1.4D, plus/minus tolerances must be confined to sizes (diameters, widths, etc.) of proper features of size. Applying them to anything else either doesn't suit the function and mating relationship, allows more than one interpretation, or both.
Dave, I agree that we're never going to agree on this. I know that you believe that GD&T should be applied to features that have a function and relationship, and that plus/minus should be used on features that don't. For me, it's just not that simple. I don't even agree with how you decide whether a feature has a function or not! But I agree that it's good that we can discuss various points of view.
Evan Janeshewski
Axymetrix Quality Engineering Inc.