Re: [freehaven-dev] Review: WShAnon, Berkeley '00

[dmolnar@belegost.mit.edu]

On Thu, 27 Jul 2000, Michael J Freedman wrote:

> So I just returned from the Berkeley workshop, and I thought I write
> some notes/comments/minutia/etc about the conference.

Hi Mike, 

Thanks very much for the notes. This is insanely cool. 


> Third, I think we definately gained some ideas from the conference,
> both in the modeling/ideal world sense (Roger should have more ideas
> about "Ted and his transcripts") and some more practical design and
> considerations.

This reminds me -- remember we were wondering if we could make our
definitions work out such that "indistinguishability" and "semantic
security" would be equivalent? Yevgeniy Dodis has actually reformulated
the proof of equivalence in his thesis solely in terms of "experiments"
without any reference to encryption schemes at all.

What does this mean for us? It seems to mean that we can start with
easy "indistinguishability" style definitions -- like
"given server A and server B, adversary has negligible advantage over
guessing to determine which one published a file"
and, with minimal work, get "semantic security" style definitions where
we can say "the adversary gets no information about who published a file."
At least, if we can make the definitions fit this framework, or
extend the framework to fit our definitions. 

> Proceedings of this conference in the LNCS series.  If so, we 
> can submit an editted version.  I think Roger has some ideas
> for the beginning "theory" areas.  A friend at ZKS also
> suggested that our Scientology reference can be construed
> as libel - something to consider if we can resubmit.

I'm in favor of this, at least provisionally. Note that Springer
authors receive a large discount on Springer publications. :)
Not sure how I feel about the Scientology reference at this point. 

-David