Chris, you hit on something that no one has mentioned before. The double axle design has an "unheralded" advantage, which I think is the true reason why people like them so well. This advantage is that because the springs act between each lobe and the axle base, rather than one lobe against the other, there is a greater tendency for the lobes to remain at equal retraction, even though one may be careless in placement by "pushing" the cam into the crack, rather than "placing" it.
In reply to:
I think waht BD is refering to in the section you mention, is that when fully extended the cam is at the limit of it's rotation. Look to the section on how the cams hold the applied force, in the first part of the paper.
Someone brought up the difference between 1 axle SLCDs and the Camalots. One of the biggest differences is in the range of cam rotation. The Camalot uses the axles to stop the cams from overrotating. Limiting the largest part (largest radius) of the cam from going above the axles (holding the cam vertically). This is regardless of where the other cams are.
Single axle cams have the "cam stops" on the upper part of the cams. So the (single axle) cams are limited by the postion of the opposite cam. The largest part of the cam can rotate until it is completely opposite the opposing cam. The cam stop moves with the opposite cam.
Check this with some SLCD cams in your hand. Hold 1 cam compressed and then move the opposing cam through the full possible range.
Both single and double axle SLCDs are perfectly safe if used correctly. Someone also hints that the single axle design is safer than the double. The cam range issue mentioned above make the Camalot extend out to its maximum range if it should ever need to and load like a chock.