[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] email-like service atop of GNUnet?

From: Christian Grothoff
Subject: Re: [GNUnet-developers] email-like service atop of GNUnet?
Date: Tue, 06 Nov 2012 09:51:11 +0100
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv: Gecko/20120922 Lightning/1.0b1 Icedove/3.0.11

On 11/05/2012 08:56 PM, Ivan Shmakov wrote:
>  > You've lost me here.  You've never fully explained what you're trying
>  > to do.  Based on what you write below, I guess it is some variant of
>  > pseudonymous E-mail?  If so, in what way does Mixminion not solve
>  > your problem?
>       Its documentation is somewhat scarce, but it doesn't seem to
>       implement a routing scheme to get through restricted topologies'
>       networks.  (Unlike GNUnet.)  Also to note is that I'm more
>       interested in a solution that provides strong message integrity
>       (authenticity) checking, rather than anonymity.

Ah, OK ;-).

>       I believe that contemporary e-mail has a few major issues,
>       in particular:
>       • inefficient routing and delivery — the routing is often
>         “static”, and there doesn't seem to be a way to make it adapt
>         to the location of the recipient (rather, the routing takes
>         into account the location of the recipient's mailbox server);
>         for delivery, even if the link used is 8 bit-clean, most of
>         the ASCII “control” codes have to be escaped, thus increasing
>         the overall message size (by ⅓, should Base64 be employed);
>         delivery to multiple recipients is also inefficient;
>       • spam — which forced e-mail, while initially simple, to evolve
>         into a quite complex system, with various tricks to fight
>         unsolicited mail traffic;
>       • due to the aforementioned complexity, it becomes infeasible
>         for individuals and small businesses to deploy their own
>         e-mail infrastructure, and forces them to rely on major e-mail
>         providers instead, raising privacy issues among others.

I think many will agree with this analysis of the status quo.

>       Thus, the goal is simple: a complete replacement for e-mail.
>       (And I suspect that the very same application could then replace
>       netnews, mailing lists, Web fora, and even bug trackers and
>       wikis.)
>       The basic ideas to remedy the issues above are:
>       • use public key-based identifiers as “mailbox addresses”, and
>         allow the user to “publish” the list of preferred hosts for
>         the senders to “post” the messages sent to him to; certainly,
>         it should be possible for such lists to be maintained
>         automatically, perhaps only requiring confirmation to use
>         user's private key to publish them;

What exactly is a 'host' for you? An IP address? A peer? A physical
machine (with possibly multiple/changing IPs)?  Something else?

>       • establish a loose web of trust, thus making spam largely
>         infeasible; (eventually, one's web of trust will be large
>         enough for the vast majority of incoming mail to possess a
>         signature of a “friend of a friend”);

How should the WoT database be maintained? Where is it stored? How is it
queried?  How can new users join initially (how do you plan to bootstrap
the system)? How exactly does it stop spam?

Also: did you read Martin's thesis on GADS or view the video of his
defense (both on yet?  Not quite a web-of-trust (trust is
again problematic, as you'll need to define what exactly I convey by
signing someone else's key), but a graph between authorities is part of
GADS.  Would that graph do?

>       • given that the routing can now largely be user-driven,
>         switching to, from, or between major “new mail” providers
>         would be a non-issue.
>       My guess is that both Secure Share and FreeTalk (check, e. g.,
>       [1]) are intended to solve roughly the same issues.  So,
>       perhaps, working on a similar solution on top of GNUnet routing
>       (or some other) would be a duplication of effort.  Or may be
>       not, for there still may be a lot of room for exploration.

I'm not sure SecuShare has replacing e-mail on their feature list, and
the FreeTalk discussion is again related but --- as with your above
design sketch --- all of those are still just rough ideas.  So I think a
wide exploration of the design space is at this point still justified,
especially as long as you're aware of what others are doing or have done.

My 2 cents, happy hacking!


reply via email to

[Prev in Thread] Current Thread [Next in Thread]