cable length

In any kind of cable--IC, speaker, or power--you generally want it the shortest that it takes to get between whatever points it's connecting. Of course, you'll want some slack to facilitate moving, installing, positioning, and such. But don't use 15 feet if you only need 10, for example. Increased length only adds series resistance, series inductance, and shunt capacitance to the overall circuit.

Dekay: there's no evidence of any sort of "ideal" length of cable other than the shorter the cable, the better. Period.

For analog signals, the wavelengths of 20 Hz to 20 kHz range from about 15,000 kilometers to 15 kilometers, multiplied by the cable's velocity factor (usually from 0.7 for coax to 0.95 for twin conductor). So there's not even an appreciable fraction of a wavelength involved there, and thus no need to get into transmission line theory.

For AES3 digital, the cable should have a characteristic impedance of about 110 ohms (which is approximately what most pro mic & signal cables exhibit) to match the source and load impedances of the digital devices. There's no need to match "wavelengths" or any such thing because even a reasonably close impedance match will keep reflections far below the threshold where they might corrupt data or cause jitter. Keep in mind also that the digital data is not carried directly on the cables but is encoded onto a carrier. This further helps isolate the data from effects of cabling to maintain its integrity, and it also frees us from worrying about polarity in AES3 interconnects.

And as Stevemj said, power cables are not in the signal path and will not affect audio performance. As long as the audio devices get the power they need, there will be no effect on the sound. Low-powered devices such as CD/DVD players, DACs, etc., have regulated power supplies that can tolerate a fairly wide range of AC voltage without losing any performance.