Well, it wouldn't be the first time someone had the right idea decades before the world was ready. (No sarcasm; fully serious.)
Though you'd want to open it up to people beyond open source developers. And it's possible the idea of using the "up/down arrows" that everyone is familiar with may still be worthwhile; it's been too long since I've seen the advogato interface to remember if they had that. And finally, a "likeability" metric that was less mathematical and a bit more rough & ready might be a good thing; mathematical constructs are really good at maintaining the properties you design in, but are often extremely fragile. I'd want an algorithm with some knobs I can twist, not a single algorithm. But by all means read the papers in question; why rediscover those things the hard way?
Though you'd want to open it up to people beyond open source developers. And it's possible the idea of using the "up/down arrows" that everyone is familiar with may still be worthwhile; it's been too long since I've seen the advogato interface to remember if they had that. And finally, a "likeability" metric that was less mathematical and a bit more rough & ready might be a good thing; mathematical constructs are really good at maintaining the properties you design in, but are often extremely fragile. I'd want an algorithm with some knobs I can twist, not a single algorithm. But by all means read the papers in question; why rediscover those things the hard way?