The great thing about this country is that everyone is free to do things the way that they see fit. Regarding installing studs the way you describe, go right ahead and keep doing it your way. Hardened bolts have a finite number of times that they can be stretched/released (why do you think they break?). If you want to start your bolts out that way then that's your decision. Second, the while you can pull a stud on that way, the threads are not designed for that purpose. I'll continue to install them my way and if asked I will recommend the same. You are free to do the same thing and the OP can decide what way he wants to do it.
Everyone is happy then, right?
ETA: One thing I just thought about is that while this method can work with a used hub/rotor, I bet it won't with a new rotor hat or hub. In fact, I would bet that in most cases the stud would break due to the forces involved. The splines from the old studs have already cut the knurl pattern into the old part(s), placing less pulling forces on the new stud/threads than when installing new studs into a new rotor/hub. In this case, the stud splines have to cut a new path into the new part(s), which are made out of hardened steel. The OP is in this position with the new rotor hat and I would not recommend your preferred method. Even then, sometimes a BFH and drift are not enough for the job and that's when it's time to hit up the local machinist if you don't have a press. In this case, the rotor hat cross-section is thin enough that the BFH method should work fine.
Regarding there being a tool for drawing a stud into place, as a mechanic over the years I have seen many brainstorms implemented in tool form from mechanics. Many of them are unnecessary or detrimental to performing a good job while some of them are good ideas. At one time or another every good mechanic has made a specialized tool to work on an application where a regular tool won't work. Some are good ideas, some are not. This tool might be great for a shop that wants a job done quick so they can get on the next one but I'm the kind of guy who wants the job done right because it's my car.
The lives of my family depend on good repairs in a job like this, I'll take the time to do it right.
This is one of the instances where your intuition is misleading you. The phenomena you're referring to is fatigue, and for a given high-grade steel one can stress the bolt to very near its yielding point (i.e. stressed to it's 'proof load') tens of thousands of times before failure is induced. The proof load for a 12mm stud is around 18,000-20,000lbf (equivalent to a 9-10ton press) which should be plenty to draw in any stud. For a 14mm stud the proof load is around 25,000-27,000lbf.
You are perceiving that your method is the 'right way' to do things because you've never seen a failure. I would wager that there are people out there that have never had a stud break by drawing it in with a nut, but have seen issues when hammering in studs. Neither cases provide any justifiable, tangible, proof of being the 'right' way of doing things.
Here is my reasoning behind my opinion, back by some real technical understanding. I may not have wrenched on boats and cars for the last 30 years, but I am a mechanical engineer who has designed and built cars from scratch, raced auto-x and endurance, and have spent the last 8 years or so wrenching on cars.
About the studs... Wheel studs are hardened steel, and hardened steel gains strength at the sacrifice of ductility (ability to deform without cracking). So I would choose to draw [12mm] studs in with a nut because the threads and shank are
certified to carry 9-10tons of force through their threads, shank, and cap without any significant detriment to longevity. The torque required to generate this force on a 12mm stud is around 170ft-lbs, which is quite a lot of torque for a small stud all things considered. The reason I would choose not to use a hammer is that the hardened steel stud is less ductile (more brittle) than your average steel and thus more prone to cracking. A brass drift does help avoid cracking because it will deform before the steel stud, but it just doesn't play to the (literal) strengths of a hardened steel stud. Both are probably fine methods of installing a stud, but I prefer to use a nut because IMO the likelihood of over-torquing a wheel stud is very low.
Stud failures are (I believe) most commonly found at the root of the threads because of bending stresses being concentrated by the threads near the bottom of a stud. Bending stresses are usually a result of poor torquing procedures. Though I will certainly acknowledge that gross over-torquing of studs can most certainly lead to straight-up tensile failures.
About melting loctite... I'll address the faulty example of the cylinder sleeve that you gave first. "Thermal abuse" isn't really a thing... there will be a lot of heat (as in joules) passed through the areas around a cylinder sleeve, but the temperatures seen by the outer surface of a cylinder sleeve are relatively low due to the engine's coolant. The outside surface of the cylinder wall should never go above 230-260F (pressure cap rating dependent). If these temperatures were surpassed there would be spot-boiling in the cooling system, a rare occurrence usually only present in serious track cars. Thus loctite used on cylinder sleeves would stay well within its operating range.
Regarding loctite melting around wheel studs. The heat generated in a brake rotor is dissipated in two ways: convection to the air flowing around the rotor, and through conduction into the wheel and hub. Now, yes, the wheel does make a really good heat sink. This is a reason why many race cars and performance cars use aluminum wheels - aluminum has great thermal conductivity. But a brake rotor is much more thermally-coupled to the hub than to the wheel because they share a much larger mating surface area. Thus the hub face will be at nearly the same temperature as the rotor hat. Also note that the studs themselves don't have to be at 450F, just the hub itself, the outer perimeter of the loctite melting is enough. The studs themselves are actually somewhat thermally isolated from the hub and wheel because of the splines and lug nut, those don't provide a lot of real mating surface area, and loctite is a lousy conductor of heat compared to metals.
Here is a really interesting article wherein an S2000 was driven up to 60-70mph and then braked down to 0. After 4-5 runs of this the brake rotor top-hat is around 560F, and after 10 cycles one can clearly see the top hats reaching nearly 770F. To me, this is VERY strong evidence that the hub could easily reach 450F. Of course this is quite an extreme amount of brake heat, most of us aren't running out there and going 0-60-0 ten times in a row. But, coming down a long steep hill or mountain, spirited driving, etc. could lead to a similar situation IMO. More to the point I have seen "dark-blue" to "blue" discoloration on street car brake rotors which
indicate rotor temperatures north of 550F. This isn't a common thing to notice since regular use of the brakes will simply grind this coloration off (the layer is very thin) but it is anecdotal evidence of temperatures similar to the MotoIQ article on street cars' rotors.
Also note that in these MotoIQ tests that the wheel is considerably cooler than the brake rotor, the rotor surface itself is considerably cooler than the brake top hat, and the outer tip of the stud is still quite cool despite the brakes and wheels being considerably warmer. This firmly disproves the assertion you made that the wheel, studs, hub, etc. are all at the same temperature. They can very clearly be different temperatures, by hundreds of degrees.