but the above circuit won't work.
Are you sure?
I'm not.
You need a pulse generator, with a variable duty cycle, to control the output voltage, and some means of feedback.
Like peter-h wrote this won't work as there is no oscillator to be seen.
Not sure, but can't it be the generator?
First both transistors are on. When current in L reaches the saturation voltage at MOS drain rises (provided supply source and MOS isn't too strong) switching both transistors off (VR1 is the positive feedback). When whole energy from L will be transferred through diode voltage at drain jumps down starting cycle from beginning.
Not being sure I can expect that VR1 can be a feedback regulating the output voltage.
Or you can vary the pulse frequency but the control loop gets messy.
It is the other way around.
Control loop happens to have stability problems when you want frequency being constant and only duty cycle being variable. But when you don't care about frequency things are much simpler.
And this circuit (if working) is certainly not constant frequency converter.
In 2014 I've selected to use in our products LM5017 from TI converters new family (developed in 2012). It is able to step down from 100V down to 3V3 using simple choke (without taps). I like its COT (Constant On Time) mode. Taking advantage of the fact that it doesn't keep frequency constant COT is free from feedback stability problems. I specially like its very fast reaction on load changes. COT changes duty cycle in single pulse time while constant frequency, current mode converters need 5..10 pulses to notice that duty cycle change is needed. I also like that it works in exactly the same mode being not loaded (internal key not simulate diode and if load is too small energy from L is transferred back to source).
I assume that if someone emphasizes that he is a beginner than learning is in front of getting ready to use complete module. If this circuit will work (which I do not rule out) the lesson will be learned. If not the lesson also will be learned