I've been thinking about this lately, too.
So, what I do now is:
a) teach recovery to a glide descent,
b) recover to a glide descent, then recover to S&L (as per ex 8)
c) recover to a glide, then recover to a climb
d) recover with minimum height loss, to a climb (which I think is what the examiner will be looking for).
The rationale here is to show in a) the pure aerodynamic recovery with change in A of A, back to the known (glide recovery to S&L), THEN introduce the ideal in d). Too many people on reval. flights shove in the power before the A of A has reduced, causing a pitch up back into the stall, with all kinds of yawing going on 'cos they don't adjust rudder against the power slipstream changes. I'm looking eventually with a 'straight' stall, for a constant heading through the manoeuvre.
TOO