No, he said "Rotating line...like the second hand of a clock." So, think of it this way: A scope, in it's common (i.e. usual) mode, generates a horizontal line. How does it do that? With a sawtooth voltage on the X axis that has just enough amplitude to sweep the beam the full width of the display area (or typically a large percentage of it).
Now, if you put the scope in X-Y mode, and do the same: apply a sawtooth wave of the proper amplitude (call it V) to the X input (with the Y input at zero), again you will get a horizontal line. If you add a similar waveform (same frequency, phase and amplitude) to the Y input, you'll get a line that is at a 45 degree angle. And if you remove the waveform from the X input (so it stays at 0 Volts), then you'll get vertical line. The frequency of the waveform is uncritical and only needs to be high enough to engage the persistence of the screen phosphor so there is no flicker --on my scope that's around a minimum of 1kHz.
To make this more like a clock second hand, move the line to the center of the screen (i.e. translate the coordinates to the center of the screen), by calling (V/2,V/2) the origin and vary the amplitudes of the two waveforms (each running a around 1kHz or above) at a rate that will rotate the line at the speed that you want, from V/2 to V back to V/2 then to 0 and back to V/2 with a 90 degree lag between the two. The function of the variance is sinusoidal.
So, that would look like this:
* Starting at 12 o'Clock:
Y is swept [by a sawtooth wave] from V/2 to V at 1kHz (i.e. a 1kHz sawtooth wave with amplitude extents of V/2 to V)
X is at V/2 with a zero amplitude.
* 1 o'Clock
Y has an amplitude extent of V/2 to V/2 + (V/2)(sin60)
X has an amplitude extent of V/2 to V/2 + (V/2)(cos60)
* 2 o'Clock
Y has an amplitude extent of V/2 to V/2 + (V/2)(sin30)
X has an amplitude extent of V/2 to V/2 + (V/2)(cos30)
* 3 o'Clock
Y is at V/2 with a zero amplitude.
X has an amplitude extent of V/2 to V
* 4 o'Clock
Y has an amplitude extent of V/2 to V/2 - (V/2)(sin30)
X has an amplitude extent of V/2 to V/2 + (V/2)(cos30)
and so on...to make this smooth, simply vary the extents smoothly, rather than in steps (as shown, above). Note: I only showed it in clock face steps to get the idea across of how the waveforms change