On Tue, Jul 11, 2006 at 09:13:03AM +0100, Michael Rogers wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Matthew Toseland wrote: > > We can grant a darknet node > > more tokens than an opennet node; we can increase the maximum > > bucket/queue size, we can increase the likelihood of allocating tokens > > to a darknet node, or even not fill the darknet buckets at all until > > the opennet ones are full. > > Increasing the probability sounds better than increasing the bucket size > - - the bucket size is only meant to determine how large a burst of > requests we can accommodate. > > Even if we don't end up mixing darknet and opennet, it would be useful > to be able to give different priorities to different peers.
Perhaps so. Is it sustainable? If we give fewer tokens to a node then they will have to give fewer tokens to us; there is misrouting here, what sort of impact will it have? > > > If anyone has a better > > non-gimmick idea for encouraging people to use the darknet even if they > > have opennet connections already, please give me it. > > I disagree with Ian's characterisation of useful applications as > gimmicks. There are some applications (eg instant messaging, chatrooms, > blogs, voip, shared folders/workspaces) that match the darknet model > very well. But they're not part of Freenet's core goals, and therefore they are worthless, or so Ian tells me. > A darknet that supports these applications will provide > significantly better privacy for users of these applications than the > current centralised solutions. This is not about using gimmicks to trick > people into using Freenet: it's about giving people more secure, private > versions of the applications they already use. > > I'm not suggesting that Freenet should *only* support single-hop > communication - I'm all in favour of building a general-purpose > anonymous communication network, and I understand the arguments for > keeping it as general-purpose as possible - but in my opinion we need > the darknet model for security, and the darknet model happens to be a > good match for a pretty broad and popular class of applications, so why > not allow the users to benefit from that? Because it's not part of the core freenet goals. It's a red herring, a gimmick. Or so Ian would tell you. Convince him, or at least convince Oskar. Or do it yourself; even Ian probably wouldn't revert such patches. Unfortunately you in particular we need to be working on simulations, because we need the new load limiting algorithm ASAP, but somebody else could work on it; but not me. > > Cheers, > Michael -- 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/20060711/3a6fc512/attachment.pgp>
