[Enigmail] Societal and technical obstacles to widely-adopted encryption

[Robert J. Hansen]

Michael Kane wrote:
> That's what I'm talking about re: Skype.  Skype (if we assume their 
> encryption is good) is a very easy-to-use and secure communications 
> system.  You don't assume that the people using Skype are "privacy
> nuts".

Yes.  But this gets back into the architectural argument.  People don't
need to be privacy enthusiasts in order to keep their conversations on
Skype encrypted because Skype incorporates that from the very ground
level of the protocol.  It's fundamental to the protocol at a very base
level.

With email, encryption is a voluntary add-on above and beyond the normal
application stack.  Bolted-on solutions will always have architectural
fragility, and that will always be an obstacle to widespread use.

The problem here is not OpenPGP, not GnuPG, not Enigmail.  The real
problem is SMTP, which is just about the worst standard I can imagine
for email, with the possible exceptions of every other standard I've
seen proposed.

> In my opinion, ease-of-use is the key to any technology flourishing.

This is not borne out by research.  If ease of use was the key to
technology flourishing, everyone would be using Macs.  Many, many
factors go into the adoption of technologies, ease of use being only one
of them.  Chrome and features are two other factors which appear to have
more bearing on technology flourishing than whether that technology is
easy to use.

> There are certainly examples of hard-to-use technologies that win (at
>  least for a time).  (See MySpace, Windows and eBay)

Given that none of your examples have yet to be dethroned, it's hard to
argue "at least for a time".  It is the nature of all things to have
their day in the sun and then pass.  For your hypothesis to be true,
things that are hard to use would have to be supplanted with things
which are easier to use--and honestly, I don't see that happening.
Windows 3.1 did not get supplanted by MacOS System 7; it got supplanted
by Windows 95.  Windows 95 did not get supplanted by MacOS 9; it was
supplanted by Windows XP.  XP is not being supplanted by OS X; it is
being supplanted by Vista.

> But if email encryption was as easy to use as Skype, the social
> stigma would not be an issue.

I think you mean to say "if encryption was incorporated into the SMTP
stack, social stigma would not be an issue".  Probably so.  As it's
phrased, though, I strongly disagree.

> The key is hiding all the hard stuff from the non-techies.  Even the
> crazy-looking signature block -- that could certainly be collapsed
> from view.

Only if non-OpenPGP-aware clients knew enough about OpenPGP to collapse
the block.  However, this presents a logical paradox: we're positing
that non-OpenPGP-aware clients are really OpenPGP-aware.

Making OpenPGP easier for Enigmail users is an admirable goal, but it is
a very small part of an overall goal of "easy email encryption".  If you
want easy email encryption, then you have to make it easy _even for
people who don't care and don't have it installed_.

This is why I think S/MIME has a huge, huge head start over OpenPGP.
Don't get me wrong, I prefer OpenPGP (for technical reasons beyond the
scope of this email), but S/MIME is doing much better in terms of
adoption rates and transparency.

> But the only way to get to that promised land is to make it SUUUUPER 
> easy for people to set up.  Rob, you mentioned that Ubuntu +
> Thunderbird + Enigmail already handles some of this.  But it
> definitely doesn't pass the grandmother test.

Yes, because SMTP does not have it incorporated at a fundamental level.

> * a single installer

Already have it.  Ubuntu, Portable Thunderbird, etc...

> * a simple walkthrough for creating a new email account

Already have it in Thunderbird.

> * a simple walkthrough for accessing your existing mail account

Already have it in Thunderbird.

> * encryption by default, so that UNencrypted mail would seem unusual
> and uncomfortable.

Phenomenally stupid.  OpenPGP-encrypted email accounts for well under
.01% of email traffic.  (Yes, it's that small.)  If you make unencrypted
mail unusual and uncomfortable, you're going to make it unusual and
uncomfortable for 99.99% of mail traffic.  No one will bother with it.
It's too unusual and uncomfortable.  Those few souls who do bother with
it will get looked at funny, because after all, why are they using
something so unusual and uncomfortable?

> * email signatures by default that advertise the privacy and easy
> setup of this free software package, so that recipients could adopt
> it, too

People already hate banner ads, clickthroughs and basically all kinds of
internet advertising in general.  Bad idea.

> * automatically updating address books so keys are shared easily when
>  new people install the software, making it easier to opt to send 
> encrypted mail.

This is called the PKI problem and it is /phenomenally/ difficult to do
well.