On Tue, May 30, 2006 at 09:21:55AM -0700, Ian Clarke wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > On 30 May 2006, at 08:22, Matthew Toseland wrote: > > >On Mon, May 29, 2006 at 11:09:53PM -0700, Ian Clarke wrote: > > >I did hope to be able to > >distribute invitation files, which would be a once-only invite > >allowing > >you to connect even if you don't have a node yet. The problem is > >that if > >both parties are NATted (and they usually are), this won't work > >because > >they need to know each others' IP addresses. So I'm not sure how we > >can > >make it easy to connect. > > ARKs may help here.
No, they won't. You have to have a two way exchange of references, if the newbie is behind a NAT and is in fact a newbie. > > >My original objection stands to FCP connection management: 99% of the > >usage of FCP connection management will be for grotesque hacks which > >produce bogus opennets without the right topology. > > That demonstrates a rather condescending view of client authors. This demonstrates established fact. Have a look at the numerous key exchange boards on Frost, the 2+ ubernodes, #freenet-refs, the clients complaining about overload because they used a script to add 300 (disconnected) peers... > Anyway, as I have said before, the point is moot because people are > *already* implementing "grotesque hacks" using fproxy, we need to > provide a better alternative, and the best way to do that is to > provide proper support in FCP, which will include concrete advice > about how to ensure the kind of clustering required for small world. So they can more efficiently implement grotesque hacks? No thanks. > > >The remaining 1% will > >be for fproxy replacements and the like. But if you have any specific > >examples of things that you could do other than automated > >connection to > >random published peers, I am willing to listen to that. > > *Any* attempt to implement a more user-friendly mechanism for > establishing connections, which may in many cases be platform > specific, will require exposure of connection management > functionality via FCP. Options include email-based solutions, IRC > plugins, AIM plugins, and all sorts of other things. Why can't we do these things ourselves? How exactly do we make it easier to get connected? > > Generally, when you are trying to stop your users from doing things > where there is an entirely legitimate reason for them to do it, on > the basis that they can't be trusted, you are on the losing side of > the argument (and you are in bad company, this is the kind of > attitude that gives rise to awful ideas like DRM). It's easy enough for them to do it anyway. What we don't want to do is give our implicit approval to the practice by making it easier. > > Ian. -- 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/20060531/e9c17cab/attachment.pgp>
