photopassjapan - 1:39 pm on Dec 11, 2006 (gmt 0)

I guess there's just isn't enough CPU to compute the entire index again based on this filter, so instead, G checks pages it indexes, reindexes, and can match up to others, based on a couple of easy to spot parameters. I mean there have to be some flags that will trigger the algo to even bother to check a page against another one. And so they could at least guess which one that "other one" should be.

Title and meta may be a trigger for example. If either is a match in the index, they check the contents, if not, they don't. It's just not possible to check every page against every page on the net based on some search-for-this-sting type of script.

...

It's so not possible.