Up to this time there was no need to differenciate output from input in the mixer, because input and output levels were the same because there was no realtime effects, but now we have something like a chain, so levels entering the effects chain are not the same than levels outputting from that chain.
Now I have a bunch of realtime level sensitive effects in the chain, like compression and I cant use the mixer to adjust levels because it influenciates the sound coming out the chain.
So the solution would be to keep input and output levels separate and handle both from the mixer by adding an output tab in order to control pre and post levels independently.