If the train identification principle gains momentum, I think we can expect Märklin to take that step sooner or later. But will they then go for the Littfinski/Helmo RFID principle or the IR principle used by Uhlenbrock Lissy system? The RFID principle has inevitable technical advantages (cheap non-powered units in the rolling stock, no sensitivity to ambient light, no need for line of sight) while the reader side can be built much cheaper for an IR solution. I have built an IR system myself, and the component cost is no more than 3-4 Euro per sensor.

If Märklin adopts train identification at all, I guess they will use RFID, because an IR sensor is hard to install in a C-track. But on the other hand, they could design a special C-track with built in IR detector. Whatever they do, it will probably be expensive!

It would also be very interesting to know the reliability for the existing systems. Both principles can theoretically give very high reliability, but how is it in real life ? I have managed to get my home made IR system extremely reliable, but I went through a number of design iterations with not-so-reliable versions before I got there.

The s88 feedback bus is essentially a synchronous serial channel with a fixed frame length of 496 bits. There is nothing in the serial protocol that demands that each single bit should correspond to one specific contact. You can take a bunch of bits in the frame and code whatever you like, as long as you can get the SW in the other end to understand it. So there are obviously lots of possibilities!