I have given this topic a bit more thought as it has developed.
*** I suppose i had better just say that i am testing a fully screened valve with no chance of looking at the internal structure for this exercise. ***
With heaters one thing that has not been looked at yet is the increase in resistance as the valve heats up to the correct temperature.
This could be tested for different envelope types and periods of manufacture in order to estimate the voltage by measuring the temperature related increase in resistance as a percentage for a number of known valves.
Start with 0.5 volts as some heaters can be as low as this in things like hearing aids.
Keep the results in a table
Cathode location would be next.
First make sure that a high ohms range does not pick anything up from any other pins with the negative probe connected to the heater with the heater still on. This proves that it is indirectly heated.
If this is the case then connect a 100K resistor to the high output of a signal generator and connect the other end to the high end of a scope probe and common the earths.
With the heater off connect the earth to the heater pins and dab it on each pin and note how much it is shunted by the cathode to heater capacitance. The greatest shunting effect will identify the cathode pin. This is not required if it was identified as directly heated.
If there is no easy connection to the external screen just slip a foil covering over the envelope and connect this to the low side if the test probe and repeat the shunting effect test and note the two pins that show the greatest shunting effect in order to identify the screen and anode.
If there is an obvious screen on the envelope just measure it on an ohms range to find the pin.
Next with the same 100K resistor apply DC voltage in place of the signal generator to it and turn on the heater and measure the shunting effect starting at about 60 volts with a DVM.
No shunting effect identifies that there is an internal screen present and that the other electrode that showed a strong shunting effect is the anode and transferring the probe to it should confirm this.
Now just measure all the other pins and hopefully they will just act like dropper and some idea of grid spacing between cathode and anode can be estimated from the voltages using a DVM.
Any pins that did show a shunting effect that fail to show voltages on this test can be tested again as a separate active element on a common cathode with the already identified pins noted and eliminated from the remainder of the test.
If the internal structure can be seen some short cuts can be taken.
At least this should give a good idea of the pin out.