Backward propagation as a first step makes sense — it's the lowest friction option and at least gives nodes visibility into what's happening. Starting with upfront fees right away would probably slow down BOLT12 adoption at the worst possible time.
The real question is whether "low-grade degradation" from an attacker under backward propagation is tolerable enough to buy time for a proper fee-based solution later, or if it just becomes the permanent state nobody bothers to fix.
Backward propagation as a first step makes sense — it's the lowest friction option and at least gives nodes visibility into what's happening. Starting with upfront fees right away would probably slow down BOLT12 adoption at the worst possible time.
The real question is whether "low-grade degradation" from an attacker under backward propagation is tolerable enough to buy time for a proper fee-based solution later, or if it just becomes the permanent state nobody bothers to fix.