Terry is right, that there is no specific "rotary" axis in the UCCNC.

All axis are handled as linear axis which does not mean that you can't do rotary work, you can:

https://www.youtube.com/watch?v=Ox4SrYt-qEYAnd non of your points 1. or 2. is the case.

When you code G1 Y10 Z1 B720 F120, then all axes will move syncronised and all axes will reach the target point the same time and the feedrate F120 is on the vectorial sum of all axes which is a virtual space of YZB axes in this case, but it could be upto a 6 dimensions space of XYZABC in case you code all axis in once.

It's hard to visualise this, because the space in this world has 3 dimensions only, but if you think into it I'm sure you will understand.

To do so called specific "rotary axes" support a software could have 2 options (The UCCNC does not support any of these at the moment):

1.) To support rotary axis in a way when it is defined how to compensate the movement distance on that radial axis. So, that the feedrate will be calculated on the surface of the cylinder around the rotary axis.

E.g. to define that the A axis is rotary and it's center line is at Z=0 and so the software will then know the geometry, the circumference of the axis, because it then knows that the radius is defined by the Z axis.

2.) To support kinematics and so the user let the software know the properties of the machine, around what lines are the rotary axes are bending around in the workspace and also how the axes are connecting to eachother. A good example is a rotary BC table where the table can bend to one direction and there is a table on the top of it which can rotate around.

So, then the software knows the kinematics of the machine and so it could make the required matrix transformations to execute different motions, e.g. linear interpolation and orientation commanded moves.

Another example is a robotic arm where the joints connect to eachother and so doing a linear interpolation with set feedrate which requires matrix transformations on the geometry to calculate how the joints have to be rotated to do the commanded linear or other movements with the set feedrate.

Ofcourse the first point is much simpler to achive than the second. I know, because I've studied robotic arms a long time ago when I was in college, we even had to do exam in which calculations like that had to be done on paper. It's not easy math I can tell you, especially when you have to do it on paper.