Longer is better?

Both AI, @mitch2, and those who agree with them are correct. Anyone who remains skeptical or chooses not to believe this can conduct the experiments themselves. Below is my actual experience.

I had an SMSL SU-1 DAC, which accepts digital input via USB, Toslink, and RCA (coaxial S/PDIF). When I used a non-75-ohm RCA cable, the sound became distorted, paused intermittently, and suffered from dropouts. After switching to a true 75-ohm RCA cable, even one as short as 3 feet, the sound was completely stable—on par with USB.

This demonstrates that the primary issue is impedance mismatch, to which the SMSL SU-1 is particularly sensitive. When the cable and connectors are properly impedance-matched, signal reflections do not appear to be an issue even with varying cable lengths.

I then tested a couple of USB (Type-A to Type-C) cables ranging from 6 inches to 6 feet. In all cases, the sound remained stable, and the sound quality did not vary with cable length.

Any cable is an additional (passive) electrical circuit.
Any additional circuit degrades some of the electrical qualities or the source. Never improves.
Shorter cables degrade less.
That’s physics, nothing else.
 

Yes, we don’t want to complicate the problem, but we also should not oversimplify it. Again, digital RCA cables are more susceptible to impedance mismatch and reflection issues than USB. We need to address impedance matching first, rather than being overly concerned with cable length.