Glengara posted at SEW asking whether Fantomaster's suggestion that on-page SEO was being ignored, to the detriment of rankings, was accurate. Several good responses illicited even more questions on the subject.

I, Brian in particular raised two excellent points:

Google has repeatedly tried to raise the bar on the worth of links, particularly by the creation of a core group of documents regarded as of especial importance, with all other recommendations following from them. Hilltop, LocalRank, and Trustrank, are all famous Google algorithms founded on this key issue of determining links of most worth from the rest...

Certainly it would be unwise to ignore on-page factors. While links can be essential for targeting competitive rankings, the overall aim of any SEO method is to capture targeted traffic.

AndrewGoodman also contributed brilliantly:

There seem to be several on-page and off-page factors alike that seem to be under-discussed on these forums, maybe because people like to believe in old myths (Google's algo is mostly about PageRank and always will be), and because many SEO's are overconfident in what they already know and are afraid to admit want they don't know or can't control.

Google is, as Brian notes, a link-based search engine, and the other major search engines have largely followed suit. While on-page optimization is typically regarded as important, it doesn't get nearly the attention in the SEO world that links do.

Perhaps, as Andrew suggests, the pendulum is swinging away from link based algorithms somewhat. Certainly the recent advancements in processing power would suggest that text analysis for both quality and relevance could have dramatically improved in the last 2 years. It's possible that 3-5 years from now, links will be only a small part of the ranking algos, as the "highest quality" writing on a particular search subject is rewarded with top positions.

Right now, I'm picturing thousands of SEOs signing up for expository and technical writing courses around the world.