On Tue, Apr 04, 2006 at 01:39:28PM +0100, Michael Rogers wrote:
> >We could really
> >do with some local and semi-local stuff e.g. instant messaging (as an
> >extra incentive for people to add their friends, for example).
> 
> The more I've thought about f2f networks the more important local 
> communication has seemed. Global search and routing are problems from 
> "light" networks that are not only harder to implement in "dark" 
> networks, but make less sense there. Rather than taking the internet or 
> traditional p2p networks as our model, I think we should be looking at 
> im, irc, livejournal and myspace. "Friend management" and communication 
> with friends are a significant part of these systems; browsing and 
> searching are relatively unimportant.

Well, should we be adding people from seeing their blogs? That is
precisely how LiveJournal works, but it is dangerous from a darknet
perspective...
> 
> Friend management seems to have at least three aspects: presence, 
> visibility, and introductions. Presence means being able to see when 
> your friends are online; fproxy already offers that. Visibility means 
> being able to see who your friends are friends with, and to choose which 
> friends are visible to which others. At the moment freenet takes the 
> most secure but least social approach: friends are invisible to one 
> another. But in order to encourage the third aspect of friend 
> management, introductions, it might be useful to allow people to make 
> selected friends visible to selected other friends.

Well, introductions are probably the better way to do this. Send
somebody a message "you might want to connect to X; connect to him if
you know him".
> 
> Introductions are important for a friend-to-friend network because they 
> allow people to make new connections without out-of-band communication. 
> The introducer must be trusted not to perform a man-in-the-middle 
> attack, but by definition the introducer is someone the other two 
> parties already trust. (For extra security, the new friends should be 
> encouraged to confirm one another's keys out-of-band at the earliest 
> opportunity.)

Yeah, we should allow introductions, but we should put some careful
warnings in...
> 
> How do we make introductions as painless as possible?
> 
> 1) Allow users to make selected friends visible to selected other 
> friends, and make it easy to browse your friends' friends, possibly by 
> hosting myspace-style profile pages for your friends.

Not a good idea IMHO. Allow users to introduce a specific friend to a
specific friend.
> 
> 2) Provide a link on the profile page for requesting an introduction to 
> a friend's friend - the friend's friend will be sent a request, and if 
> the request is approved, the introducer will give each friend the other 
> friend's ref (including the public key).

No. We should not encourage people to expose their friends to their
friends, except by way of specific introductions.
> 
> 3) Automatically confirm your friends' public keys through other mutual 
> friends, to make MITM attacks more difficult. For example if Bob 
> introduces Alice to Carol, and Alice and Carol subsequently discover 
> that they both know Dave, they should automatically exchange refs 
> through Dave as well, making it harder for Bob to perform a MITM attack.

I don't get it. Bob could have made up a new node with a new key. We
have to do some sort of out of band verification... if only by asking
people to confirm introductions out of band.
> 
> 4) Provide a "web of trust" page showing whose ref was obtained or 
> confirmed through whom, and which refs have been verified face-to-face.
> 
> 5) Allow users to host local blogs and chatrooms that are only 
> accessible to their friends. Again, provide an easy interface for 
> friends to request and approve introductions.

That's a good idea.
> 
> Anyway I'm sorry if this sounds like a long list of feature requests. 
> This is the sort of interface I've been planning for my own vapourware 
> f2f network and I thought some of the ideas might also apply to freenet. 
> Any thoughts?
> 
> 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/20060404/ad4aa147/attachment.pgp>

Reply via email to