How about a gold case, nice-looking connectors a shiny display and impressive knobs?
To follow this topic a bit further, the Audio-gd M9 appears to use relays for attenuation, and has 100 steps of attenuation. If I am understanding all of this correctly (not likely) that means that there are 100 relays, and one of them is opened for each step, allowing the signal to flow to one of 100 different resistor circuits, thus varying the output signal level. So the components in the signal path include the open relay and whatever resistors are in the selected circuit. The relay controller is not technically in the signal chain, since it is simply telling the relays to open and close according to external commands. And of course this is duplicated for each channel. Right?
Alternatively, if I am understanding the LDR system correctly, the the only thing in the Tortuga signal chain is the LDR, which varies its resistance based on how much light hits it. I assume that means that the LDRs are shut off from all external stray light in the critical wavelengths, and I'm guessing that those wavelengths are well into the optical part of the spectrum, otherwise you would get significant drift due to changes in temperature introducing stray light...??? In any case, the light has to be carefully controlled to ensure proper attenuation, because the LDR response is not perfectly linear with changes in the light. Based on earlier discussions in this thread, the response of each individual LDR is also unique, so you would have to either map each LDR's response and then teach the controller how much light to shine on that particular resistor to get a particular level of resistance, while hoping that it never changes (fat chance) or instead put some sort of feedback loop in to measure the actual output level and adjust the light accordingly in real time. With the latter, the chain is not perfectly pristine any more, and is definitely not as simple as the relay system, which would seem to me to be a bit more predictable, as long as one assumes the resistors are stable in their performance over time.
So how much of this did I get right?
You are pretty well right on. The LDRs can drift over time, but it's not any more significant than a normal resistor. There's also the benefit of updating the micro-controller at a later date to ensure tight calibration. No feedback loops for the reason you stated.