By thoroughly discussing the design choices and trade off's with regard to existing lnurl implementations, backwards compatibility, added complexity, effect on incentives with regard to centralisation pressures and so forth. Move fast and break things is not a good philosophy in protocol development imo.