On 22 November 2013 15:31, Solomon Peachy <[email protected]> wrote: > On Fri, Nov 22, 2013 at 09:50:43AM +0000, Dave Cridland wrote: >> However, I stress - the point, to me, of the 4th January test is not to cut >> connections to Google, or send some Message, or anything else along those >> lines. > > Oh, I completely agree. > >> The point is to see what happens, accepting there will be some disruption, >> and accepting that we may have to re-examine what we think is achievable >> here. > > Strong end-to-end encryption is the ideal situation, and we need to make > sure there are no technical impediments to achieving this. Therefore, I > will gladly participate on this (and future) interop/test days.
The problem with this "ideal" is that it still leaks metadata. Also presence is never likely to be encrypted end-to-end. You get the idea. While there are certainly protocols that could allow you to not trust your server and your contact's server, they come with their own different limitations that I don't think make it feasible for XMPP to ever go in that direction. For this reason I think that, as much as end-to-end encryption is desirable, we do need to trust our servers to some extent and that servers should respect that trust and encrypt their links. Regards, Matthew
