--- Toad <[EMAIL PROTECTED]> wrote: > On Mon, Aug 11, 2003 at 01:15:36PM -0400, Nick Tarleton wrote: > > On Monday 11 August 2003 09:57 am, Some Guy wrote: > > > Ok, maybe someone has a good arguement why this type > > > of senerio is improbable. Does anyone have data to > > > suggest that something like this can/does happen? Is > > > there a way to gather this data? Perhaps I could > > > write some simple simulations based on the old code > > > from the freenet paper, where nodes have something > > > like "tastes". > > > > > > My fear is that learning based on keys could be > > > sidetracked by learning based on "taste". While there > > > might be advantages like collecting together > > > "like-minded" nodes, my guess is that if something > > > like this does happen it might hurt the routing. > > > > > > I hope my posts haven't been too pie in the sky > > > lately, I'd really like to help out with the routing; > > > it facinates (and frustrates) me. > > Since everything is hashed and encrypted up the wazoo, files on any given > > topic will be more or less evenly distributed throughout the keyspace, and > > there's no way for a node to tell what a file passing through it contains or > > is called. > > It's better than that. Nodes won't specialize in doggy pics, hopefully, > because nodes handle requests for unrelated content that happens to be > in the same area of the hash, from requests from the rest of the > network.
I admit it's just conjecture I have no proof it would happen. I was just thinking "hey why might the routing work in a simulation, but not work in real life." That was back in August, I think before NGR. You argue on the website that if two adversarial nodes post and request junk data to DoSA the system, they will wind up being close to each other in the network. If that's true it seems logical that people with the same "taste" would wind up near each other too. Nodes of with dog fans will may tend to find nodes of other dog fans better than nodes of cat fans. The problem might get worse if most of the cat fans are in China, because the latency would make it more unlikely dog nodes and cat nodes will connect. I have no proof it would happen. It seems like it shouldn't happen if NGR is working well, but it also seems like if it were happening NGR, wouldn't work. Hmmm, maybe it's best to imagin starting with 2 completely seperated dog and cat networks. If you connected a few links across, would they grow together or not? __________________________________________________________________ Gesendet von Yahoo! Mail - http://mail.yahoo.de Logos und Klingelt�ne f�rs Handy bei http://sms.yahoo.de _______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
