Sorry Chris, but I think you're wrong. I rarely like to deal in such black and white terms, but while the idea of the aircraft moving relative to a package of air is a fine way to think about navigation and the like, you really can't forget that an aircraft's inertia comed from its groundspeed (well, technically we need to get even more complicated than that, but bear with me!). An aircraft's airspeed will determine the aerodynamic forces available for it to turn, while its groundspeed will determine its momentum, and thus rate of turn, turn radius, etc.
Take for example an aeroplane flying at 60 knots IAS with a 60 knot tailwind. Fairly obviously, it will have a groundspeed of 120 knots, and so 120 knots worth of momentum. If that same aeroplane turns into wind, with sufficient added power to compensate for the extra induced drag, then, assuming a relatively instantaneous turn, it would suddenly find itself with 120 knots groundspeed into a 60 knot headwind, resulting in 180 knots airspeed. The large resultant increase in drag would rapidly reduce the airspeed to 60 knots, resulting in zero groundspeed. Of course, this drag doesn't wait until the aircraft is pointing directly into wind to take effect; it will begin to bite the moment the aircraft turns slightly from the wind, and is therefore easily disguised by induced drag effects in the turn. It will be more noticeable in heavier aircraft, as JF's Harrier example shows.
Chris - your statement about sink/climb rates is also suspect - you've neglected to consider the effect of runway length. With a sufficiently long runway a turnback would still be viable even if your sink rate was higher than your climb rate, as you won't be aiming to arrive back at the same point you got airborne from. Feel a bit like I'm having a personal dig at you, promise I'm not, sorry!
All of which is fairly well off topic, apologies!
Turnbacks are still a currency item for RAF single engine FW QFIs. Tucano turnbacks have already been thoroughly described here. Hawk turnbacks are only routinely practiced to runways other than the reciprocal of the departure runway, with an additional consideration being that the approach end barrier of the runway in use (ie the one likely to be needed to stop a heavyweight aeroplane landing downwind at high speed on the reciprocal of the departure runway) is generally electrically isolated, and would therefore likely be unavailable. Tutor turnbacks are very similar to the Bulldog variant described - 80 knots and an immediate 45 deg AoB turn into any crosswind being the key parameters.
For my money I'd always get the 180 done first, rather than worry about jinking away from the crosswind for separation from the centreline. I'd rather sort out the line up after at least pointing myself in the general direction of the airfield crash services. That is, howver, just my thinking, and jinking first may well be a better plan in aeroplanes with steep climb angles, to leave more runway available for the downwind landing by tracking away from the runway initially. So many variables!