Do you mean no integrated circuit regulator? If you are trying to get a constant 5.1V from a 6-9V supply then, by definition, you are creating a voltage regulator.
I'm not sure what you mean. It works with or without a load.
A better circuit is attached, which adjusts the voltage depending on the current through the load so that the load always gets the same voltage.
That said, you won't get 5V1/200mA output from 6V using the topologies shown - there simply is not enough voltage headroom. You'll have to use a more intelligent approach than the emitter follower and current sense as each of those drop at least 0.6V.
What has you learnt in class? Perhaps you have learnt about operational amplifiers? You can design one using transistors and this can be used to regulate the current without such a large voltage drop.
That's what the lower transistor is doing. As the current increases, that transistor starts to turn on which will reduce the drive to the upper transistor, thereby limiting the current. Feel free to simulate it.
I got 3.3V across the lm317 and 1.24V across R1 at the knee of IR2 for an overhead of 4.5596V for the current limiter set up, add to that 2V across R2 and you get the 6.5 volts I incorrectly referred to as overhead, good catch.
If you are getting under 3 across the LM317 and I'm getting 3.31516V, it's probably the LM317 model, I use the LM317psp from regulator.lib.
Kinarfi