Andrew, I agree with disagreeing.
In
perpendicular_snap.png there is a second perpendicular point on the far side of the white Circle.
See feature request
FS@2621
Not an issue: There are 2 tangential points for an RZ outside the (full) Circle, no when inside.
But the user remark was not limited to Circle shapes ...
If the white shape was an Ellipse there might be
2-4 perpendicular points and
0-2 tangential points.
Max 2/4 perpendicular points simply depending on whether the RZ is inside the evolute of the Ellipse or not.
This evolute may extend far outside the Ellipse shape itself.
Not an issue: There are 2 tangential points for an RZ outside the (full) Ellipse, no when inside.
Typically only
one perpendicular solution is returned based on
getVectorTo(), the near perpendicular point.
This method is not really accurate for Ellipses and there are a few documented false solutions.
I am testing alternatives + A tool for all (2-4) normal points is as good as finished as we speak.
Detail, a Circle is a special case of an Ellipse and the evolute of a Circle is the center, a point with no distinct enclosed area.
Then when RZ = center there are infinite perpendicular points or no distinct solutions.
If the 'other' entity is an Arc and if that is open in the direction of RZ we also miss the
far-side perpendicular point.

- PerpendicularToArc.png (604.04 KiB) Viewed 29086 times
For Splines/Polylines this may be the 'Closest Simple Shape' from a virtual explosion.
Typically an RArc shape, again only the near-side perpendicular point is returned.
Which 'closest simple shape' that is considered can be unclear depending on the zoom state.
Regards,
CVH