I agree. This spec looks very under developed yet and I'm not sure it
actually addresses the key issue which prevents more universal adoption of
end-to-end encrypted email.

The problem is and remains one of key management and humans, which becomes
even more difficult because it is trying to retro-fit encryption to a
protocol  which has no support for it.

The real challenge for specs like this is that they mean nothing unless a
majority of mail clients support them. Getting them added is going to be
extremely difficult - near impossible. It is more likely we will just see
people move to different comms channels which are secure rather than trying
to retro fit traditional email.  A bigger problem is that their spec for
level 1 only deals with users using a single mail client. I'm not sure in
this day of multiple devices this will be sufficient. The reason I moved to
an imap based setup is that I regularly use 3+ different mail clients on 4+
different devices. I need to be able to access my email from all of these
devices and this spec will fail to provide that.  Level 2 looks like where
this functionality will be targeted. However, the problem is that level 1
may not get the uptake/momentum needed to get to level 2.

Note that I'm not meaning to disparage the effort - it is a good/meaningful
effort. However, I think it needs to mature a fair bit before any real
implementation/support can be added to existing mail clients. It is likely
efforts like this are what is needed to work out a better solution, but
this spec so far seems to lack some meat. It needs to provide more detail
on exactly what the key problems are which prevent automatic end-to-end
encryption of email and how this spec will address those problems.


Tim


On 28 August 2017 at 19:41, Robert Pluim <rpluim@gmail.com> wrote:

> Rajeev Narang <rajeevn@riseup.net> writes:
>
> > Is anyone working on supporting Autocrypt in emacs.
> >
> >   https://autocrypt.org/en/latest/level1.html
> >
>
> I don't think so. I didn't see any rationale on that page for why we
> need yet another way of transferring keys and specifying encryption
> options in email headers, perhaps you could explain why Autocrypt
> would be a good thing?
>
> Regards
>
> Robert
>
>
>


-- 
regards,

Tim

--
Tim Cross