A bit of context here might be useful. I started this thread with the primary objective of finding a fluid head that would handle my 25x100s, the answer to this has gravitated to a Manfrotto type fluid head with counterbalance (501hdv). I almost always run binoculars AND a scope at the same time and my thoughts have evolved into how to further use this type of fluid head to simplify my observing experience. For me having similar mount characteristics for both my binoculars and scope will do just that.
An L plate with a saddle is a good solution if more than one scope will be mounted. For me, I had the flat plate stock available and I was only looking to mount the one small scope in that manner. I kept it simple.
I don't see why the 501HDV wouldn't work well as the basis of a "side saddle" setup. It has a bit lower profile and wide set pivot points to spread the load. I do prefer the smoother, finer range of clutch action of the 3063, though, since the CB spring isn't necessary. With an L plate you'd have to provide enough clearance to reach the side tension knob of the 501; with the 3063 the clutch lever can be located on the opposite side. Certainly you want the least offset for best results.
Alas, I have restricted myself to one wife but have been unable to control myself when it comes to telescopes thus the desire to be able to mount various sizes. You are correct that the 501hdv will require more offset than the 3063. My hope would be that the wider stance, dual pivots and lower profile of the 501hdv will compensate for the increased offset.
I'd proceed with caution on the video heads with side torque. Photo heads are designed for pitch, roll, and yaw movement. Video heads really aren't designed for heavy torques on the roll axis - most video cameras have their centers of mass near the centerline. When I really loosen the yaw friction (az axis) on the 501HDV, there is already noticable movement in the roll axis. That is where I feel it would weaken the fastest under side torque. You can, of course, double L and use counterweight on the other L to help guesstimate balance at the cost of total weight. There is no real way of knowing beforehand what they can handle because that use is very far off spec and you can't unlock in that direction to test balance. Knowing when you went too far though ... ?!
I agree with you that side torque or roll, if not immediately an issue, will become one. The double L or U saddle plate would allow for balancing. This may be an advantage for me, I use a navigational array (laser, telrad, 70mm RACI) which can mount on the opposite side from the scope to help lateral balance. I can minimize guesstimating balance by balancing the rig off the head where it would be free to roll.
Total weight becomes an issue of course, especially if dead weight is needed to balance for roll. One solution is a heavier head, i.e. 503hdv or larger. I would prototype first and try it out on the 501 and move up if necessary. As you say "far off spec", but is that not half the fun