The "optical" version of the barn-pole problem

We present diagrams and simple calculations for the apparent (i.e. photographable) length of a moving ruler skimming the observer's position, under three different classes of model. Special relativity's predictions in this particular situation are the root-product average of the two more basic first-order predictions generated by simple "propagation timelag" arguments. We find that special relativity can legally predict either a photographable Lorentz contraction or a photographable Lorentz expansion in the "centred" ruler, depending on whether our camera is at the ruler's "apparent" or "official" centre.