Servo wheel/arm rotation in a horizontal plane (as with typical servo mount) definitely causes some side-to-side motion of the pushrod/wishbone configuration. When I wiggle the servo end of the pushrod (unattached to servo) from side to side, even a little, I can see differential movement in the elevator halves. Some guys say there is no visible effect to the plane in flight, but if it's visible on the bench, I don't want it. So I've been side-mounting elevator servos for the last half-dozen planes or more.

The Central DEPS system is pretty good, but I make something pretty similar with Darroll Cady pushrods, just as effective and a bit less expensive. These dual pushrods do need one or two bracing/damping points along the mid-tail area, and on the Javelin you'd have to cut into the balsa sheeted bottom to do it, then re-sheet and re-cover. I ususally strip all the covering off as my first step with an ARF. Sorta contrary to the general idea of ARFing, but not a big deal to anyone with kit-building experience.

The vertical arc in servo output for a side-mounted servo will introduce some non-linearity of control movement relative to TX stick movement, but that domain is already full of non-linearity; but at least the differential between elevator halves is gone. And besides, once you're you're flying, you move the sticks according to what your eye sees the plane doing, not according to the geometry of where the stick is.