Flutter is definitely based on the airflow velocity (i.e. TAS) because it happens when the airflow has just the right velocity to stimulate/reinforce a natural resonance in the piece of the airframe in question.
This reinforcement is not strongly dependent on the dynamic pressure i.e. IMHO it will happen at a given TAS and not be much dependent on the altitude.
Is their any possibility that the pilot had inadequate supply of Oxygen and their recollection might not be reliable? Quote
That would be my guess…
I agree with:
The aeroplane behaves exactly the same way as in lower levels, but the controls feel very “soft” and control response is slower than usual. Altitude loss during recovery is higher, but well within 1000ft. If you don’t recognise the stalled condition and let high (or low) angles of pitch develop, it can take some time to come back to normal pitch. Maybe this is what happened to the CitationJet described in Neil’s posting above. But it does not really explain what the TB21 pilot experienced.Quote
The absolute ceiling will be defined (though probably not stated in the FM) as the altitude at which the plane is flying at maximum available power (turbo or not), and no longer able to climb. Not able to climb means that the pilot will be on the verge of a stall. Mis handled in any way, and it will stall, and maybe spin, if mis handled more. The recovery will be the same technique, but likely to require more effort, precision of control, and time.
During Mogas testing I got a normally aspirated, carburetted Cessna 185 to 20,800 feet (with oxygen), and was in this condition. I did not stall it, but the stall horn was warning. I did “zoom” the plane to just over 21,000 feet, but it would not stay there. Some non pressurized planes do state a “maximum operating altitude” I presume for handling.
Aside from crossing the Himalayas, I cannot think of any operational reason to have a GA aircraft that high, or in that restrictive aerodynamic condition.
I also recall that the aircraft was sluggish at high altitude, FL400 or so in the sim (Cessna CJ3). Stall recovery was the same though.
The article about flutter and TAS is very interesting. What I was thinking was whether control wheel feel and weight varies with IAS or TAS. However the point about air density dampening flutter probably answers the question. It seems that control effectiveness and feel will vary with IAS and not TAS.
Aside from crossing the Himalayas, I cannot think of any operational reason to have a GA aircraft that high
I would agree that one would not normally fly that way for fun, not least because anywhere near the ceiling needs “best power” (about 120F ROP) which gives you at least 10% less MPG than “best economy” (peak EGT or LOP).
But there is a fair bit of convective weather here in Europe where the tops are c. FL190-200 with a fair consistency, and being able to overfly that without buying a turbocharged aircraft is quite useful. Today is a good example. I scrapped a flight Shoreham-Dortmund
The real TS activity is in the southern part of the stuff
but you don’t want to fly in IMC in any of it really.
To be honest, I didn’t see much real TAS/CAS explanation in that Australian Flying article.
He says early on that air density and TAS are both important, but ignores air density from that moment on. Why?
And shirts on a washing line are not even remotely elastic! 
I dug around a little and found this much less readable, but probably more reliable, NACA technical note 4197
In equations 9 to 16, they have the flutter speed ‘Vf’ as a TAS, appearing squared on the left,
and the air density ‘rho’ appearing in the denominator on the right.
The speed of sound ‘a’ is also there, but appears on both sides of the equation, (they lose it on one side later).
The essence of the equations is that this kind of flutter occurs when dynamic pressure equals a function of some stiffness/inertial/Aspect ratio stuff.
This means a “flutter line” could indeed be painted on the ASI.
So unless there is some other kind of flutter they did not know about in the 50s, I would vote for CAS.
