tedster - 11:37 pm on Aug 21, 2012 (gmt 0)
I don't see how that approach scale, because a "site" cannot be reviewed in isolation. Its individual pages would need to be reviewed with reference to a particular query phrase.
At the scale of today's (and tomorrow's) web, it looks to me like algorithms are a necessity. Penguin is an early attempt at a new kind of algorithm (the ground was first broken by Panda) and it's only a beginning, I'm sure.