Bob: The classical view treats light as a wave, so therefore it has a front, which travels at, well, the speed of light. The wave is seen as having amplitude, frequency and possibly phase. Light is of course an electromagnetic wave, therefore the wave consists of two parts, a magnetic field (A/m) and an electric field (V/m), which are at right angles to each other. Unlike other waves an electromagnetic wave does not need a medium to travel in, it can exist in a vacuum. Note the units of A/m and V/m, the ratio is in units of resistance; this is the impedance of free space, about 377 ohms, which also dictates the relative strength of the magnetic and electric fields. All this applies in the far field, ie, greater than half a wavelength from the source of the wave. In the near field, much less than half a wavelength, the wave is pretty much magnetic only, no electric field. In between the electric field gets it act together and we end up with the classical electromagnetic wave. The definitions of near and far field are a bit imprecise, as you might expect, since there are no sharp transitions. For light the wavelength is pretty short, less than a micron, so the distinction between near and far field is largely academic in everyday life. At lower frequencies this is not the case. A few years ago I designed a radio system running at 30Hz. Why so low? Because it had to work through 1" of cast iron, and over a range of a few metres. The wavelength at 30Hz is about 10000km, so clearly the radio was operating in the near field, essentially magnetic induction as there would be no electric field to speak of.
However, as Jason describes, light has wave/particle duality. Sometimes the behaviour of light can only be explained by treating it as particles, ie, photons, rather than a wave. Due to the Heisenberg uncertainty principle we can never exactly know both the momentum and position of a photon, so if we know it's momentum we don't know exactly where it is, and vice versa.
For the SDRAM issues mentioned the tracks and copper planes on the PCB are designed such that the signals actually travel as an electromagnetic wave in the material of the board. The speed of travel is slower than in free space, due to the relative permittivity of the board material. This value isn't known particularly accurately, 10% may be, so the argument about picoseconds mentioned was largely theoretical.
Regards,
Andrew