Hmm, will have to check my whizz wheel tonight...
As for your counter-argument, you say:
However, if increase the pressure to sea level assuming ISA and then take it back up past the station elevation assuming ISA (which is what the altimeter does), this makes little difference
Yes, that's true. But if the temperature is not ISA, you can increase the pressure to "sea level", but this is not the true sea level.
Example:
Station elevation is 300 feet. Assume 30 feet per millibar at ISA temperature, and an average temperature of ISA-10 degrees.
Measure the QFE as 1003mb.
Now we can calculate the QNH as being 1013mb [1013 + (300/30)].
Next, dig a hole, 300' deep. (Or alternatively, assume the station is on the top of a cliff.) Go down to sea level, and measure the pressure. It will not be 1013mb. You will reach 1013mb 12 feet above sea level (calculated using the rule of thumb of 4% per 10 degrees, since I don't have a whizz wheel with me at work). The pressure at sea level, therefore, will be slightly higher. (This pressure is the QFF.)
With QFF set on your altimeter, I agree completely with what you're saying. But with QNH set, what you are calculating is not your height above sea level, but your height above this 12 foot point. If you want to get the correct answer, you have to bear in mind that the only point at which the altimeter will read the correct altitude is at 300' - the altitude of the station. Again, let's give it some numbers... assuming you are flying at an indicated altitude of 900', with QNH set. According to you, the error is (using my rule of thumb again) 4% of 900, 36 feet - your true altitude is 564'. According to me, the error is 4% of 600, 24 feet - true altitude of 576'. Drawing a diagram of this situation will prove that 576' is, in fact, the correct answer. 564' would be the correct answer only if the question stated that you were flying on the QFF - but when does anyone ever fly on the QFF? Note also that I could not have got to my answer without knowing the station elevation of 300'.
FFF
-----------
EDIT - have just thought of an even easier way of explaining why DFC can't be right (thanks to a response to your duplicate question on the Rotorheads thread, which I just stumbled across).
Assume you are at 12,000', as per the original question. You have QNH set. But you are actually on the ground, at an airfield with elevation of 12,000'. The temperature error
must be zero in this case, because QNH is the value which you set to get your altimeter to correctly read airfield elevation when on the airfield. Using DFC's theory would have your altimeter reading some huge temperature error, despite the fact that you're on the ground. You can not do temperature correction problems without konwing the elevation of the station.... QED!
FFF
--------------