On the other hand, when a mini DisplayPort cable is plugged into a Thunderbolt port, the Thunderbolt port can resolve the signal from the DisplayPort signal. This is actually what is meant by back compatibility.
You wouldn't use a TB cable between an old mDP only mac and the display.
Conceptually, it isn't really a big problem to route the internal DP traffic in a pure "pass through" mode if:
1. there is no TB network present on either TB socket.
2. there is some DP device chirping on the other end of wire looking for someone to talk to.
The controller has to pick up and forward DP traffic if it the leaf node in a TB daisy chain. This is just a TB chain length of zero an the signal is inside.
However, it is should also be simpler to only look for DP displays in search of a host. As opposed to hosts looking for a display. For example if someone hooks up two DP cables, each with video cards, to two TB ports. Which card is the display suppose to talk to? the TB controller has to implement a DP switch? If it sticks to matching displays to single GPUs it is simpler.
If it is bi-direction handshake and has a 1-to-1 pass thru mode, the disconnect is why Apple doesn't say something about what works. They'll claim things like AirDrop works (which it doesn't if you have older implementation of 802.11 ) , forgets to mention FCPX won't open FCP7 projects, reimplements FileValut to be something new and your encrypt home may not mount ...... but in the case where something does work.... they go quiet as a church mouse. Highly uncharacteristic.
A simple footnote next to the system requirements that mentions this wider audience for the product is more likely to appear on the marketing/sales pages.