On Tue, Sep 20, 2005 at 02:55:55PM +0100, Ian Clarke wrote: > On 20 Sep 2005, at 14:08, Matthew Toseland wrote: > >On Tue, Sep 20, 2005 at 12:58:44PM +0100, Ian Clarke wrote: > >>On 20 Sep 2005, at 11:33, Matthew Toseland wrote: > >>>>Well, if that would truly be the topology then the alternative is > >>>>"clusters of isolated dark nodes", which is worse? > >>> > >>>There would be no real reason to grow the darknet, that's the > >>>point. If > >>>the only way to connect (easily) is by growing the darknet, it will > >>>grow. > >>> > >> > >>So you propose to force people to run darknet nodes even though they > >>might be quite satisfied to use the opennet? I don't believe in > >>forcing users to do things against their will. > >> > > > >Eh? I don't understand. If they want to use the opennet, they can use > >the opennet. > > Yeah, but then they can't be part of the darknet. You are saying to > people: "I'm sorry, you can only connect to people you trust, you > aren't allowed to connect to strangers". If the user wants to > connect to strangers, and those strangers are happy to connect to > them, then it is futile for us to try to prevent it.
They can connect to whomever they like. I object to there being one network with both harvestable path folding and darknet topologies. I don't know if the routing would work, or how we could make it work, but I also object to it on strategic grounds. > > >>>I don't believe people would make the effort to grow the darknet if > >>>they > >>>are connected by open nodes. And furthermore, if they are > >>>connected by > >>>open nodes, it tells us nothing whatsoever about the viability of a > >>>fully dark network. > >>> > >> > >>People get a choice. If people chose to leave their nodes open, then > >>so be it. It isn't our place to force people to do one thing or the > >>other. > > > >In which case the whole experiment will have been totally > >pointless, and > >there will be NOTHING to build on in the future, because we won't have > >actually prototyped the globally scalable darknet. > > Perhaps according to your definition of "darknet" we won't, but my > definition of "darknet" includes the choice to connect to strangers > if the user is willing to take that risk. Of course you can connect to strangers. But using path folding to do it means that it is *no longer a pure darknet*. The segment of it which has path folding will probably be the vast majority of the nodes. We are no longer testing darknet routing, we are testing opennet routing. > > What you don't seem to realise is that we don't get to choose whether > or not people will connect to each-other indiscriminately, many > people will regardless of what we say to them. We will see "Freenet > matchmaking" websites set up that will probably ruin the network's > topology as they will have no regard for the requirements of a small > world network. They can use the opennet. And it will only ruin the network in those areas of the network whose occupants allow it to happen i.e. who put their nodes up. And it won't work, because the topology is broken, so the people who used the matchmaker will go back to the opennet. > > The best option is to offer people the choice, and if they want an > open node, then at least we can ensure that it won't screw up the > small world topology. > > Sure, we might not know for sure whether it could have worked in a > "pure" trusted-link network, but who cares so long as it works in a > realistic scenario which is a mixture of open and dark nodes? It's not a realistic scenario in China, Saudi, or any other hazardous environment. And it does not tell us anything we do not already know, because the proportion of unharvestable darknet nodes will be low. We're not just trying to build a working network here, we also want to validate it. Even automated matchmakers are harder to harvest than path folding - partly because only idiots will use them. > > Ian. -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so.
signature.asc
Description: Digital signature
_______________________________________________ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
