...oye vey...
Your reasoning for DCC is, unfortunately, a bit...well, You've come up with a myth to perpetuate a belief system. Or in otherwords, you've started with a result and then rationalized how that answer best fits the question. Hardly scientific at all!!
The fact of the matter is, the motor inside the locomotive still runs off a DC voltage. This does not change regardless of where you put the control source. Ever. Well, ok, maybe we'll convert to AC power someday, though I'm very skeptical of this EVERY happening.
Now in the old days, which for me was 1989, we learned how to get that sort of performance you discuss with a basic powerpack and a finely tuned and calibrated wrist. That's what we learned how to do with those hours spent running around the carpet oval: we calibrated our instincts to produce accurate visial train movement. But of course, you want the train to do Everything Automatically and all you want to do is turn the knob, and you don't want to learn how your locomtoive naturally runs, you want it to do all that automatically for you too. So this means we don't do that anymore.
So we have more devices in use to control the locomotive performance and you put a box in the locomotive. However, your "box" does not have to be IN the locomotive in order to provide similar operation. It could be in a box under the layout, or it could include a box with a rhetostat that then talks to the box under the layout. Either way, the location of the box won't bother the DC motor one bot - it may make the perpherial features a little interesting, but this is where you turn on your brain box and remember what engineering capacity is available to us.
All the ideas you discuss in this article as being only available in DCC are equally viable in DC if you're willing to spend the time to electrically engineer the solution. BEMF, for example, could just as easily be done using a couple capacitors in the motor circuit, so that any changes in direction require an electronic dump followed by an electronic gain to get moving again. If you needed to make modifications to operational characteristics, a couple linear potentiometers with screw adjust in the circuit would give you an intermediary point at which you could augment how the locomotive responds to voltage, how much voltage charges the capacitors, etc.
The problems of constant lighting were also solved years ago. Directional lighting, as complex as it looks, turns into a diode circuit that directs current based upon which direction the axle is turning. Sound might seem like an issue, until you realize the volume control is accomplished with a slide switch, driver timing is accomplished with a mechanical or optical cam, and individual sounds have also been figured out by the manufacturers as well.
So the argument you have given is indeed a sieve without any wire between the holes. Yes, water can fall out of a bucket without a lid, but has DCC all but made you forget about gravity which holds it there in the first place?
For those who are still operating DC, there are still layouts where more than one locomotive will never be not only unnecessary, but quite simply not feasible due to the limited arrangements of the trackwork. Despite all the magical qualities of DCC, Trains still don't pass through each other, jump over each other, or step asside so another train can go past them. And as long as this is true, DC is still a very inexpensive option for a very large number of people, particularly given the number of DC locomotives available in the universe, the short supply of space so many have, and the insatiable zeal so many have for collecting. A calibrated wrist alone is a very CHEAP, particularly when one figures the saving per unit provides enough cash to purchase another 1 to 5 similar DC locomotives - depending of course on the price of the decoder and the price of the locomotive.
The greatest selling point about DCC at this point in time is that it is the easiest and cheapest way to put all your bells and whistles into one unit, set it on the track, and forget it. It requires no thinking whatsoever to do it, either - you just put the decoder in, nowadays, and you;re done. Most people, I reckon, never get any further inside their decoders beyond setting the road number. In short, it comes down to a matter of convienience.
At this point DCC also provides forward compatibility with the future, a furture you rightly state is pretty inevitable - in otherwords, you can take the unit to the much larger club layout, friend's layout, or your future layout, and run it there too.
But this arguement that a motor will behave differently on DC power from one intermediary device versus DC from another intermediary device, well, it's simply absurd!