I didn't want to jump on this discussion, but I was curious to find out if things are really this simple.

I am mostly familiar with the typical network computer telecommunications protocols and associated schemes. Since I didn't initially believe Altec's statement that bits are simply dropped, I decided to poke around the Internet to learn if something similar to the type of error correction that goes on in the telecommunication world is done with digital audio signals. I found quite a bit about AC-3, PCM and DSD. For instance, there is an extensive amount of error correction done in the CD player itself. However, I saw very little mention of transmission errors and much more concern about encoding/decoding "quantization" errors, which seem to be more audible.
Here is the Wikipedia entry on PCM which talks a little bit about error correction schemes.

In computer telecommunications, let's take Ethernet for example, we use schemes like manchester encoding to ensure the integrity of the transmission at a very low level. As Altec points out, things are apparently much simpler in the digital audio transmissions we are dealing with here - this surprised me initially, but it makes sense since this is a point-to-point asynchronous exchange (a "bitstream", now I get it). It looks like all the work is left to the encoding/decoding process to deal with data integrity at the payload level. In which case, all things being equal, I don't personally think the type of cable matters at all, any more than the type of optical transmitter/receiver does - but that's my opinion, you are entitled to yours. This could bring us into a whole new argument as to which encoding scheme is better, but that's another thread. I found this paper (in pdf format) which summarizes some of the important concerns in one single document.

In any case, my ears cannot tell the difference and this whole 1s and 0s talk makes me glad that I left my computer science background behind and went with Information Systems for my PhD.