> The only (possible) solution I have come up with to the problem is the sexy > but probably impossibly impractical idea that the answer to the question: > "What can find the closest key fast having only a closeness relation between > them?" is, of course, Freenet - so the DataStore ADT should be modeled as a > Freenet microcosmos within each node (with each node unit in the microcosmos > having a datastore small enough for linear searches). If somebody has a better > idea I would be VERY interested. Sorry if I'm a bit behind here, but whats wrong with using a hash table that uses the routing key as the hash, and then just check previous and next hash entry for closeness?
I'm a lazy bastard and dont check the implementation but as java uses int's in hashtables and routing keys are alot longer shouldnt the highest bits of the routing key be used in the hash table so that close entries that point to the same direction dont take excess space? Then again local entries are typically close to each other and should therefore use the lowest bits as the hash.. :-P --typo _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
