Is a float expression specifying the distance from the geometry instance around which to calculate the buffer.
Is a float expression specifying the tolerance of the buffer distance.
Tolerance refers to the maximum variation in the ideal buffer distance for the returned linear approximation.
For example, the ideal buffer distance of a point is a circle, but this must be approximated by a polygon. The smaller the tolerance, the more points the polygon will have, which increases the complexity of the result, but decreases the error.
Is a bit specifying whether the tolerance value is relative or absolute. If 'TRUE' or 1, then tolerance is relative and is calculated as the product of the tolerance parameter and the diameter of the bounding box of the instance. If 'FALSE' or 0, tolerance is absolute and the tolerance value is the absolute maximum variation in the ideal buffer distance for the returned linear approximation.
A negative buffer removes all points within the given distance of the boundary of the geometry instance.
The error between the theorectical and computed buffer is max(tolerance, extents * 1.E-7) where tolerance is the value of the tolerance parameter. For more information on the extents, see geometry Data Type Method Reference.