Hi Ben,
thanks for your comments and questions. To save space, I use your numbering system below.
1) I am not sure, but in general I think the relevant binary relation is not symmetric.;-)
2) Up to the middle of page 5, the infinite frequency limit has been taken as in Fig. 3a. The result is Equation (7). The lowest order approximation of this clock is (6) and because it is translation invariant, the resulting 't' could well represent Newton's absolute time.
(7) as it stands, as you suggest, has a form of quantization in it. The quantization however just scales the unit vectors of the `spacetime' for which the 'clock' is a representative time-keeper. In taking the limit we see a link between the initial finite areas of the Compton clock, and the infinite frequency limit that shrinks those areas to zero while filling in the events on the t-axis. This leads to Minkowski spacetime with its odd signature, but by shrinking the areas to zero we have supressed quantum propagation.
The second limit is more interesting because by building in a Poisson process for events, infinite frequency is allowed, but the probabilistic weight favours a mean finite area between events. The result is the chessboard model that gives the Dirac propagator! By allowing the finite area, the multiple-history picture of the path-integral survivies and it is this that gives wave propagation. The fact that wave propagation survives in this limit and not the original is I think the most interesting feature of this sketch. It shows that SR is conventionally classical because it starts in the continuum in such a way that quantum propagation is surpressed. If you back away from the continuum, as per this clock, you see spacetime as an infinite mass limit and Dirac propagation as continuum limit in which mass is kept finite.
I do expect this picture to generalize. The above calculation shows that relatiity and quantum mechanics are siblings rather than marriage partners. We should be looking for the parent from which they emerge, and to do that we may have to forgo the initial convenience of the calculus.
(3) Yes on all counts. The scale is determined by event frequency (ultimately mass). The local dimension that characterizes clock frequency is two (spacetime area), and the oriented area (in the sense of geometric algebra) is what is averaged by the stochastic process in the chessboard limit.
(4) My background in stat mech colours my view of both relativity and quantum mechanics. In the sketch, I only mention the Poisson process that, when added to the clock, gives the chessboard model. However structural similarities between, for example, the diffusion and Schroedinger equations and the telegraph and Dirac equations are, to me, too direct to be a coincidence. The classical equations emerge from an underlying statistical mechanics, and as a result the partial derivatives involved are an approximation valid on restricted scales. I would expect the same to be true for our 'fundamental' quantum equations too.
(5) I agree. Scale dependence is common in macroscopic physics and I suspect that to resolve some of our foundational problems we shall have to be more circumspect about the continuum.