The triple-transit issue is when a reflection propagates from the Tx to the Rx, reflects back to the Tx, reflects again and hits the Rx out of phase with the "direct" signal on the wire. Triple-transit times can be important, but unless the system is really, really hosed (messed up) will not create reflections large enough to cause bit (signal) errors. Note the ASR article was originally written almost ten years ago when asynchronous DACs with good clock recovery were less common. Virtually every DAC today has circuits to isolate the DAC circuit itself from the incoming data stream so signal/clock glitches are far less important. They'd essentially have to be large enough to corrupt the incoming data, which would be very large indeed.
Of course, if the system is perfectly matched, there are no reflections, and cable length does not matter (until it is so long that its loss reduces the signal too small to be recovered as
@mansr said). In practice, given a buffered data stream as virtually every DAC does these days, it would take an unbelievably (or at least unrealistically) poor connection (large mismatch) to cause errors.
This is a big problem in my day job working with GHz data streams. Not something I worry about in my audio system.
I should update some of those old articles, but lifetime is finite...
