A naively implemented opennet will result in the same performance problems as we had on 0.5. We don't have a strong theoretical basis, load balancing would likely be different, routing flux would be much worse, and getting onto the network would, while involving less human effort, take much longer to get satisfactory performance - for exactly the same reasons it takes so long on 0.5.
THEREFORE I WILL NOT IMPLEMENT OPENNET IN THE NEAR FUTURE. True organic growth will happen on a larger scale when there is a bit more content, when the current crisis regarding backoff has been solved (this is to do with bugs, it is not the systematic problem that mrogers will be looking at over the summer), and when the tools have improved a little. There is already a certain amount of organic growth. The problem with exponential growth (growth rate proportional to size) is that it is really slow at the beginning. On Mon, May 29, 2006 at 07:09:04PM +0300, Jusa Saari wrote: > On Mon, 29 May 2006 12:29:23 +0100, Matthew Toseland wrote: > > > We're simply not ready to implement opennet. Oskar has shown no interest > > I understand that. But please understand what I'm saying: You already > _have_ an opennet. It doesn't matter if you are ready for it, it's what > there currently *is*. > > The problem is simply that an average interested person has the choice of > either getting noderefs from strangers or not getting into Freenet at all. > Given those choices, Freenet essentially operates in opennet mode right > now. > > > whatsoever in it, and without a sound theoretical basis we can't and won't > > do it. Also load balancing would probably be different and routing churn > > would likely be much more severe. In any case there is no reason why > > organic growth can't happen if Freenet provides something of value. The > > By organic, do you mean that a Freenet user invites his friends and they > invite their friends and so on ? > > That's not how it works. You surf the web and happen to freenetprojct.org, > and decide to try it. Or Slashdot publishes yet another Freenet story and > the readers decide to give it a try. Or you read an Usenet post about it, > or hear it from the IRC, or something like that. > > Usenet and IRC are close to friend telling friend -model, but web surfing > sure isn't. > > > fact that it is so small and has so little functionality that it doesn't > > is purely a short term problem. > > Actually, as far as functionality is concerned, this seems to be the best > incarnation of Freenet so far. I've carried actual discussions on Frost > boards, where I can be reasonably sure that my message will reach the > recipient. A succesfully opening freesite is a rule rather than the > exception. It works great so far - but it has little *content*. > > > On Mon, May 29, 2006 at 11:13:18AM +0300, Jusa Saari wrote: > >> On Sat, 27 May 2006 20:43:52 +0100, Matthew Toseland wrote: > >> > >> > On Sat, May 27, 2006 at 10:32:48AM -0700, Ian Clarke wrote: > >> >> People are already starting to kludge their own mechanisms for adding > >> >> connections via FProxy, for example: > >> >> > >> >> http://www.sinnerg.dotgeek.org/freenet/ > >> >> > >> >> The sooner we can support this explicitly using FCP, the less messy > >> >> this is going to be... > >> > > >> > Come on, do you really think that would help matters? We have enough > >> > people trying to hack their own broken opennets together already > >> > without making life easier for them! > >> >> > >> >> Ian. > >> > >> Current Freenet *is* an opennet. It is an opennet since most people get > >> their noderefs from total strangers in IRC. They have little choice, > >> since Freenet simply isn't large enough that an average person would > >> know someone connected to it. The whole noderef hassle simply makes life > >> more difficult for users, it does not provide any security whatsoever. > >> > >> Accept the facts and code according to them. Even if you somehow managed > >> to enforce the darkness of Freenet, you'd simply ensure that it never > >> grew beyond a very tiny inner circle. > >> > >> _______________________________________________ Tech mailing list > >> Tech at freenetproject.org > >> http://emu.freenetproject.org/cgi-bin/mailman/listinfo/tech > >> > > > _______________________________________________ > Tech mailing list > Tech at freenetproject.org > http://emu.freenetproject.org/cgi-bin/mailman/listinfo/tech > -- Matthew J Toseland - toad at amphibian.dyndns.org Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/tech/attachments/20060529/c4c40390/attachment.pgp>
