[OTR-dev] Could OTR be made more fault-tolerant?

[Ximin Luo]

On 22/06/15 12:19, Carsten Mattner wrote:
> On Sun, Jun 21, 2015 at 7:36 PM, Ximin Luo <infinity0 at pwned.gg> wrote:
>> Hopefully "real soon now" an early version will be available to test/use on
>> the MEGA chat platform.
> 
> Curious, how or why will it be in the MEGA chat platform?
> Is it an independent extension or are you collaborating with MEGA?
> 

I work for MEGA; it will be released as a standalone free open source JS library around the same time that it's linked into their web platform.

The protocol requires (but verifies, i.e. does not assume) that the transport meet certain properties but is otherwise agnostic to it, like OTR. Acks are essentially just an empty message, like in Pond, and don't require transport-level support.

- echoes the same linear order of events (join/part/messages) back to everyone
- channel members are known at all times, for each event
- the server is responsible for attempting reliable point-to-point delivery, i.e. if it claims "A is in the channel at event M" then it is responsible for actively attempting reliable delivery of M to A, even in the presence of a dodgy connection between S and A. (However, we have passive and retroactive checks that this actually happens.)

XMPP is supposed to meet these, but IRC (I believe) isn't. It's still a WIP to research in detail whether this is true in practise, and how imperfections of the transport affect the overall experience at the user level.

X

-- 
GPG: 4096R/1318EFAC5FBBDBCE
git://github.com/infinity0/pubkeys.git


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: OpenPGP digital signature
URL: <http://lists.cypherpunks.ca/pipermail/otr-dev/attachments/20150622/c61260da/attachment.pgp>