This is a cool thread. My pet peeve is when people treat digital signals like they were analog. A series of bits as clearly defined 1s and 0s doesn't have the same demands as an open-ended frequency wave. Optical cables don't need gold plated tips, and aren't effected by EMI. I can understand the reasoning for signal chain OCD if you're passing along analog signals (which are finicky and sometimes appear to abide by mystic principles) but when audiophiles apply the same rational to digital is makes them seem out of touch with technology.
True, but I've found that the really cheap optical cables that lack gold plating (even the cables I see at Wal-mart these days have gold plated ends) usually fall apart or don't stay plugged in very well.
I'm not attributing this to the gold plating, it's just something I've observed about very cheap optical cables. Usually if the manufacturer springs for a bit of gold on the tips, the cable is put together half decent.
And coax is certainly subject to EMI, although my understanding of digital transmission is that it's basically all or nothing. You go from a few errors per hour or whatever (which is typical) and experience no signal degradation, and then move up to many errors, and the signal just goes away. There is no "snow" phase where the signal is somewhat intelligible through the interference. This is based on lectures/papers I've absorbed over the years on Ethernet (and more recently QAM and ATSC) - it may be somewhat misapplied to S/PDIF, I2S, AES/EBU, etc.