I do exactly know how this works and yes on paper (Microsoft whitepapers, architecture diagrams etc) and some reports from some reputable web sites makes it sound as if it makes a 512MB system go like the clappers.
I tested ready boost over a long period on several incarnations of the OS Beta/RC and final versions of Vista on systems with different types of USB drives / readyboost sizes in different system with different amounts of RAM.
In MY testing I only noticed minimal improvements when applications loaded and did not really notice any increase on a 512MB system as there was so much disk trashing going on (paging) due to lack of system RAM. As for booting I did not really notice any difference. During the tests at one point I did look at perfmon counters however these were a little sporadic to monitorig as to gain some hard facts about % gains of readyboost over time, paging on perfmon was still high.
When 512MB test systems were upgraded to 1GB of RAM the performance of the general system did increase dramatically (as expected)
At the time I concluded that £ for £ that the better gain on a 512MB system was RAM rather than Readyboost and besides my testing it was well docuemented that systems with 1GB or more of RAM would see a diminished benifit from Readyboost.