j_t:
However, how do we effect the transition from manual flight to autoflight during the early takeoff stage in a manner which doesn't see the manual bit go outside whatever shape/dimension the trapezoid may be in a critical case with, shall we say, average pilots rather than those who might qualify for "Ace of the Base" brevets ?
No doubt we have seen that to be such a serious problem in normal ops that, in some cases, RNAV SIDS have been cancelled by the FAA at some major air carrier airports because of the variance in LNAV capture capabilities and pilot fat thumbs.
But, with an RNP AR capable airplane the transition should be smooth and comparatively easy, with LNAV track already being computed by the FMS and ready for engagement either to the FD and/or autoflight. And, if the aircraft has auto-rudder it's no different than the early stages of any liftoff after completion of the takeoff roll.
OTOH, if the pilot has to input rudder manually, and cannot remain within the splay from the DER to the linear RNP containment area then we have some pilot skills that are unsatisfactory.
I have that concern today about circumstances where pilots have to perform in "unusual" situations other than just OEI just after V1.
My premise is taking full advantage (touted earlier in this thread) of the state-of-the-art technology on more than a few airframes today. For example, with such technology an OEI containment route can be developed for Aspen Runway 33 that is devoid of rising terrain until almost to Glenwood Springs. (some 27 OEI flight-path miles.)