[liberationtech] What I've learned from Cryptocat

[Moxie Marlinspike]

On 08/06/2012 10:19 PM, frank at journalistsecurity.net wrote:
> No doubt the functional security of tools is an indispensable, essential
> concern. Ignoring any vulnerabilities is dangerous, indeed. But the
> usability of the same tools and making them accessible to
> non-technologists is just as big a concern, in my view. I know you guys
> think that many such users including Western journalists are simply
> lazy. But many, if not most of the available tools are simply not
> intuitive, or not as much as most technologists who already know how to
> use them seem to think.

I'm pretty sure we're all in agreement here: the security and privacy
tools we have today are virtually unusable.  We aren't currently
providing workable solutions for even the most basic use cases.

I don't think the problem is that we (as "technologists") don't know
this, are refusing to acknowledge it, or think you're all lazy for not
loving our terrible tools; the problem is that it's difficult to deliver on.

Here's the situation as I currently see it:

1) The crypto is the easy part!  When we were writing RedPhone, it took
me less than two days to write the ZRTP implementation from scratch.
Then we spent months trying to develop a usable interface, a
frictionless experience, and high call quality.  The project is over a
year old now, and we're still not where we want to be in terms of user
experience, but we basically haven't touched the crypto after those
first two days.

2) There are very few people actually contributing to this space right
now.  Most of the people who are interested don't come from a software
development background, and even fewer are graphic designers.

3) The resources required to produce a high quality application are
substantial.  I work on an encrypted text messaging application for
Android, called TextSecure.  If we look at what's happening in that
general communications space, the standard for user experience is set by
applications without an emphasis on security like, WhatsApp, etc.  These
other apps are the single product focus for entire companies.  It's
possible that there are ~30 engineers working on that one application,
and yet that's roughly on the same scale as the number of software
developers working seriously in the entire "liberation technology"
communications space.

So I agree with you, and it's hard, but we need to accomplish what
you're asking for while still being rigorous about delivering security
and avoiding the creation of more haystacks.  I believe we can do it,
but it's going to take time, and I think it probably means you don't get
a webapp right this second.

- moxie

-- 
http://www.thoughtcrime.org