We initially thought “Once someone sets up their probes, they aren’t going to change them”. Generally this is true but you guys are demonstrating it’s not always the case. 1-wires will always initialize in order, but adding or subtracting them will change the indexes.
We did this to make it easier for users (1-wire addresses didn’t mean much as they were discovered, not keyed in, so indexes made more sense) and to reduce communications overhead (would need to first discover the available sensors then determine which goes on which port, etc).
I suppose you could automate the discovery via a script: for example, turn on a heater, see which increases by a percentage, then assign the correct index to the correct element. This assumes each probe can be “tickled” by a device which heats or cools, but that may not always be possible in a setup sense.
Otherwise, we would need to contemplate a way to make this something you could do. The only unique identifier is the sensor address, so we would have to hand that to the user.
Another option would be to deploy more sensors than you need and just enable and use the ones you want in that current situation. Maybe not ideal but throwing it out there.