Hmmmm... seems that perhaps I learned it wrong but have been doing it right... maybe.

Never had a problem getting airborne soon enough or clearing obstacles doing it "my way", although the shortest runway I've used is probably Aeroflex-Andover (12N), which has a "generous" 1980 feet to work with and near-sea-level elevation.
I think I need to re-examine my method next time I'm in a 172 or similar... I've never taken pains to see just how much runway I used getting up and out of ground effect, and nobody's ever commented on it. I guess I probably do lift off sooner with flaps, just never noticed it. I've gotten out of the habit of observing the airspeed indicator much when taking off, other than a couple of glances during the roll to make sure it's working and that I'm approaching the book takeoff speed with plenty of runway remaining. At that moment I'm usually focused on looking out for anything that might require an abort: obstacles or engine anomalies. I pull back when it seems ready or I don't have enough room to abort and stay on the ground, whichever comes sooner. Not the best way for flying twins or jets, I guess,but OK with most light singles.
As for stability with the flaps: looks like I should also pay more attention to what's going on when practicing no-flap approaches; your comment makes perfect sense now that I think about it.
And regarding the POH note on flaps: the only Cessna POH I own is one for a K model, and sure enough, it says no flaps for a "maximum performance takeoff." But I agree with you... the runway may not really be so short that you need flaps, but when in doubt...
Interestingly, though, it also says "lift nosewheel at 60 mph" and climb at 85 mph for a normal takeoff, but does not specify a "lift nosewheel" speed for a short-field, although it does recommend a "climb speed" of 68 mph, presumably for obstacle clearance.