About 1)
The context seems to be missing. Because Toole and Olive, as far as I know, never referred to BEM simulations in their work. Their work is based on real measured data.
BEM models are usually based on ideal drivers, so the simulation results become inaccurate when the driver no longer shows piston-like behavior in reality. So this statement is true.
However, something like BEM modeling is not discussed at all in Toole's book.
About 2)
I don't know how often Toole talks about "direction" in his book, but quite often he talks about "directivity" with respect to loudspeakers. In addition to the angles, one also need the distance (since the directivity of an LS changes with distance because of the phase shift between drivers and phase shift of the secondary sound sources caused by edge diffraction).
About 3)
Regarding the frequency dependence of the dispersion, I am not so sure that it is optimal to have a fairly constant rise towards higher frequencies since this property is only a side effect of how conventional speakers are designed.
I don't recall Toole and Olive making such a specification regarding radiation patterns.
What is considered positive by Toole and Olive, however, is when sound power (SP) and predicted in-room response (PIR) slope down as evenly as possible toward high frequencies - how this is achieved is secondary.
Here, however, Sean Olive admits that it is not entirely clear exactly what this downward slope should look like (steepness of the curve), depending on the loudspeaker design.