Joe, I agree with most points, but not with the "Iref" equations. Iref is indeed the current through the reference divider, but it's supplied through the 7824 pass element, not through Rtop.
For stability, Iref is normally selected to be at least an order of magnitude above I(Q), the quiescent current sourced by the IC through the COM pin. For the 7824, typical I(Q) is 5 mA, so Iref would be set at 50 mA or so. The current supplied through Rtop is something like 1 mA, and the rest is supplied through the tube cathode and the IC.
However, that 50 mA Iref is a lot of wasted power for a 10 mA load, so a different regulator exhibiting a lower I(Q) would be better. Consider the LM317, which is essentially a "7801.25" (1.25 V regulator) having a maximum I(Q) of only 100 uA. You'd then select Iref to meet the LM317's minimum-load requirement (12 mA max), which is conveniently two orders of magnitude above I(Q).
You could probably choose reasonable fixed values intead of using a trimpot for Rtop. Consider how this circuit works: The OUT pin of the regulator IC is always negative with respect to the IN pin, and it's this negative differential that operates the control grid. The IC compensates for greater loads by increasing the conductance between its IN and OUT pins, making the control grid voltage less negative and raising the tube's conductance.
The problem is that there's a limit to how low the IC can make its IN-OUT voltage; for the 7824 it's about 2 V, and for the LM317 it's about 3 V. This voltage applied negatively to the control grid may not be low enough to allow the tube to supply the required full-load current, so the Rtop divider is in place to reduce it. However, the divider ratio must not be so low that the tube can't be controlled at the minimum-load condition when the IN-OUT differential is greatest (limited by the zener voltage).
This is probably a lot more than you wanted to know... but it's part of HAVING FUN, right?