If you are diving in the ocean (and in particular in areas that are subject to boat traffic), then you should be using a flag and float to alert boaters to your location in the water. Assuming that (a) they understand and respect what a dive flag signifies (which I know is a whole other topic of discussion), and (b) you surface at your flag, then there should not be any issues with being run over by a boat. If for some reason you are not using a flag and float (for example a drifting deco), then you should deploy an SMB to mark your position.
Either way, it should be pretty easy for boaters to be aware of your location and avoid you, so I don't see the concerns over horizontal vs vertical position on the ascent to be much of an issue in either case. You can easily be run over by a careless boater in either orientation.
That's the whole point of maintaining a horizontal position during ascent - the increased drag gives you much more control over your ascent rate. And if you have to get underwater quickly, it's much easier (as lamont suggests) to go from a horizontal position to a head-down "swimming down" position, than to do the same from a head-up position. Pike dive, pull the rear dump on your wing to get negative, and then start kicking. That's a much more effective way to submerge quickly, IMO.
Using a horizontal position to increase drag and improve ascent rate control during MOST of the ascent seems fine to me. I am however, talking about the last 10-12 feet of the ascent where I feel strongly that the horizontal position is less than optimal.
As you conceded, all the flags, floats and SMBs can be worthless when there is an idiot boater. Assuming that the ascent follows the bottom contour up into a shallow area is obviously not applicable to most all real world diving and is irrelevant to the discussion of a robust, safety-oriented protocol which can be followed in situations where boats are present.
This subject became much more than a theoretical discussion for me the day that I recovered the body of my friend who had fatal prop lacerations through the top of his skull. Looking down into his brain, as the water parted his thick hair on the ascent, left quite an impression on me. In my opinion, the most significant danger when drift diving with live boats is getting hit by a boat, not OOA, sharks etc.
I had another friend whose was on his 24 foot boat (while anchored 2 miles from an inlet in 80 feet of water) and had the dive flag up and was preparing to dive, when a 35 ft sailboat cruised (on auto-pilot) directly over his boat sinking it and requiring that he and his buddy literally leap off the boat to avoid death. Assuming that flags and smbs will always work is demonstrably a false security.
Performing a horizontal ascent, relies upon positive buoyancy and breath control to ascend. Right? So it is impossible to be listen for boats when you are exhaling. Listening for boats is a good idea, but not full proof since there are sailboats and also in some areas you can hear more than one boat.
Ascending vertically, while negative requires that the diver actively swim up. To initiate a descent the diver simply needs to stop swimming and should exhale.
In addition, the diver in a vertical position and looking up and rotating slowly (which is simple while swimming) can literally survey 360 degrees of the surface. The ability to scan the surface is less for a diver (team) who is face down. Right?
For myself, I generally hold a long speargun and have this extended vertically above my head which adds negative buoyancy (as the gun is elevated above the surface while I am still submerged) and it provides one last visual clue of the ascending diver.
And while I am going through what I do to ascend; I also usually purge the regulator for 5-10 seconds at 10 feet, which sends a huge plume of bubbles up which is often very visible to an attentive boat operator. Then I exhale, swim to maintain depth and then listen for boats before I initiate a final ascent from 10 feet.
I dont think I will argue about how long it takes to go from a buoyant horizontal position to a head-down vertical position and to begin dumping air from the BC versus a quick, negative bouyant descent since I have no empirical evidence. The other issues however, seem obvious to me.