The two patents cited above were interesting and pertinent to search engines in general, but neither patent belonged to Google or Yahoo. In delving a little further specifically into Google's patents, I discovered that they devote a great deal of effort to the customization of their search results to the predicted intent of the client as determined by his or her "current context". In a patent approved in 2010 (#7,693,825), for example, the claims section begins with, "A method of ranking article identifiers of a result set from an implicit query implied from a user's current context" ('article identifiers' = keywords), and goes on to discuss various methods of accomplishing this. Google's patents seem to indicate that they are aggressively exploring this idea of tailoring their results for specific users. I'm not sufficiently involved in SEO to devote as much time to perusing Yahoo's patents, but it would seem reasonable to think they are working along similar lines.
Thus, the same query submitted by two different users can, and should, return widely varying result sets, and the tendency is for them to become more different as time goes on. It would follow, then, that there can be nothing that resembles an absolute page ranking.
Can this really be so, or am I misreading something here? If it is so, how can an SEO expert get meaningful data with which to work?