You're being way too picky for something that 1. isn't a switching application and 2. I'm guessing you have no way of verifying (i.e., an oscilloscope, and the theoretical knowledge to relate that waveform to the circuit layout and device characteristics) whether these alternatives are actually suitable or not.
FWIW, it's not hard to do better than an IRFP260. They're ancient. Modern transistors are fully five times improved in all characteristics! This makes them much more prone to parasitic oscillation. This circuit characteristically uses linear range operation, briefly but nonetheless.
Tim