[liberationtech] What I've learned from Cryptocat

[Arturo Filastò]

On 8/6/12 7:29 PM, Moxie Marlinspike wrote:
>
> On 08/06/2012 06:59 PM, Eleanor Saitta wrote:
>> Except that with your harm mitigation, you push many potential users
>> back to plaintext, where they are guaranteed to be owned.  What
>> percentage of potential cryptocat users would the plugin version have to
>> stop from using the tool for you to accept that there was a place for
>> the non-plugin version?

I believe pure client side web apps that deploy cryptography also have
another value that is overlooked or not considered. That is that they
provide a sort of plausible deniability to the operator of a such a site.
The operator of such a system cannot trivially log the conversations of
their users, but the are required to do an *active* attack that replaces
content of the web app with malicious one. This is theoretically
detectable by the client.

I also don't think that there is that much value in having native crypto
implementations in browsers. That really does not help security at all,
it just makes things a bit more performant.
What would be really nice is a way to sign javascript code that comes
from website and disallow by default intag event binding (why should we
still be supporting sloppy coding practices?) a-la CSP.

just my 0.02€


- Art.