Gentlemen,
I do not want to discourage anyone from experimenting on there vehicles, but I am of the opinion that shaving cam towers is not the approach you want to take when running OEM cams...
First, I have run shaved cam towers "out of necessity" to compensate for a reduced base lobe cam (the Gudes). Since the 5MGE cams essentially sit atop the valvetrain components, once the base is reduced it no longer makes it down to the original position after installation... It now hovers above the valvetrain components (a slight exaggeration for effect). In an effort to restore the cam to it's original position, the cam case is shaved... This is some simple stuff here - no magic at all.
Now, as for the affect of shaving cam cases in other situations. Well, the one certainty is that you will alter the geometry or orientation of the valvetrain components one to another. This does not always result in a problem, but care does have to be exercised with a finger follower system, since there is no secure attachment point for the follower.
On the subject of increasing duration! I will tell you that there are only two ways to open a valve in this valvetrain, by pressure or by floating. Floating we definately do not want and pressure only occurs as the cam presses against the follower. With that said, "the shape of the cam is the shape of the cam!" You will not get something out of it that is not built into it. The few things that alter how a certain cam affects a valvetrain are lash and ratio. Since no one here is considering the re-engineering of the 5MGE head, neither of these are relevant subjects.
I don't think it is necessary to overstate the point of the cam influence on the valvetrain, but nor should we understate it.
The lift of the cam is dictated primarily by the height of the cam lobe, minus the diameter of the base lobe. Rocker arm ratio on the other hand affects valve lift, so it does not directly affect cam lift. A 2" piece of pipe used in place of our cam would do absolutely nothing to affect our valvetrain, but once we add a mountain peak to it or a lobe, we begin to affect valvetrain movement.
Now while the height of that lobe will determine our valve-lift, the distance from the base on one side to the base on the other side will determine our duration. If I could walk around the 360-degrees of the earth, and along the way I scaled Mount Everest, from the base on one side to the base on the other side, this would determine "duration." Duration would be the number of degrees that I had to continue walking around the earth until I would have successfully traversed the mountain. It has nothing to do with how tall the mountain is, only how broad and how long I stay atop it.
If from the opening ramp of a cam to closing ramp of the same cam is 230 rotational degrees, it is 230 rotational degrees (however you call it or position it). This simply means that the points on that cam circle where lift is initiated and ceases, will always have a duration of 230 rotational degrees.
As for lifters (HLA's) in the 5MGE valvetain. Adding shims will primarily adjust lifter preload, since these are "zero lash" engines. Even when using the shim approach over case milling (for base reduced cams), you are still simply adjusting the lifter height to a point that the initial pressure applied to the valvetrain is adequately high. It is kind of like a 5'11" guy trying to hold up a 6' ceiling (he is on his tippy-toes). But shim him up a couple of inches, and now he can apply pressure easily and comfortably within his range.
As for the affect of cam case milling on cam timing, it is not in the least bit severe. In-fact, most cars will still seem to run normally, despite the fact that there will be a slight retard in timing. While the tensioner is more than adequate to take up the additional slack in the belt, it does not account for the difference in gear orientation one to anther, since the gears are closer together (but the belt is likely being used at the same location/teeth).
Best of luck to you guys,
Carlos