diode regulator

Or you could just you two resistors as a voltage divider here, less guess work, and more precise...

use 2 resistor is diifferent than this, if the load change , the voltage is constant

Really? Because I just dropped the load to 100 ohms and it dropped to .5 volts.

It wouldn't drop that much if it was resistors... plus it is easier to match the load impedances with the source impedances to get max power out.

It's just a simple regulator , voltage is constant when load is not too small, in real circuit we use regulator chip instead

Simple? I think a voltage divider is a little more simple because the equation has already been derived for it. V_out=(R2/(R1+R2))*V_in. Which I guess you could use for this, except R2 would be the sum of the diodes, and the resistance of the diodes changes with voltage. Yes, in real life you would use a chip, like a LM7805, this used a voltage divider with a resistor a transistor and a zener diode, or you build your own using a voltage divider. When you can show me a way to predict the output voltage of this efficiently, since the diodes change resistance depending on the voltage(i.e. More voltage the less resistant they become) then I will agree it is a viable regulator.