Oskar Sandberg wrote: > > Maybe it would be useful if the node used such a list to choose one of > > its peers. That would provide a weak form of geographic closeness. > > It obviously can't base choice on which is closest, but it could possibly > weigh it into the choice. The question is how much, if at all, one can weigh > netwise closeness (really just transfer speed) to another node without hurting > the "Freenet effect" that we depend on. I can't say.
Actually what I meant was not to use a ping list for routing decisions, but for new nodes to decide which other nodes to connect with. If new nodes initially connect to nodes which are physically close, this might create a persistent correlation between freenet closeness and physical closeness. _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
