Position with one datum reference

Question

I realize this is probably a daft question, but I'm wondering why PC-DMIS allows users to callout a position tolerance of a hole (for example) to *only* a perpendicular primary datum plane using only a circle. I could be wrong but to my understanding, in the GD&T world, this would actually imply a perpendicularity callout (or in some cases, perp would actually be used instead), and we would need a cylinder for orientation in this scenario. Is it because the primary datum plane is acting as the "level" or establishing the orientation of the part/plane before checking the circle? I know that if the hole isn't perpendicular to the plane that, even with a circle, it will affect the results.

Neil Challinor · Accepted Answer

Measured circles inherit their surface vector (axis direction) from the workplane. 
 Auto-Circles either use the THEO surface vector or inherit it from the sample hits / sample feature (as long as there a 3 or more sample hits). 
 
 Strictly speaking, we shouldn't allow circles because, in most cases, they should have been measured as cylinders. However, we realised that disallowing circles would be overly restrictive for a lot of users, particular those with vison systems or those who measure thin, sheet-metal parts. The general advice is that if it is possible for you to measure cylinders rather than circles then cylinders should be the preferred choice. 
 Smashworth87 said: Is it because the primary datum plane is acting as the "level" or establishing the orientation of the part/plane before checking the circle? I know that if the hole isn't perpendicular to the plane that, even with a circle, it will affect the results. 
 To answer this part of the question, yes, the primary datum establishes the orientation of the datum reference frame (a plane locks two rotations and one translation).