Re: Nuanced points and RFC 1984

My typing is nowhere near as nuanced: slight correction below:

On 8/17/15 4:57 PM, Eliot Lear wrote:
> Harald,
>
> On 8/17/15 2:09 PM, Harald Alvestrand wrote:
> > On 08/17/2015 09:06 AM, Eliot Lear wrote:
> > > > Escrow is only useful against law-abiding people who
> > > > trust the government(s) in the first place.
> > > It's that assertion that has been repeatedly proven false with all
> > > manner of other technology.
> > {{citation needed}} - in other debates, I've heard "proof" claimed for
> > all sorts of things; sometimes it's things I believe; sometimes it turns
> > out that the "proof" is itself hugely controversial.
>
> I did provide a citation to this list on August 12th that contradicted
> what Brian wrote above:
>
> http://www.nytimes.com/2015/08/09/business/international/effects-of-petrobras-scandal-leave-brazilians-lamenting-a-lost-dream.html
>
> Here's a case where the people in question *could* have gone to some
> lengths to protect their communications but did not.  That is
> practically speaking the entire history of wiretapping.[*]  That
> doesn't mean we break our protocol suite to wire tap, but it also
> doesn't mean that we overstate the negatives when talking to others. 
> 1984 makes clear that use of escrow and other approaches has serious
> side effects that can negatively impact security.  That statement is
> good enough.[**]
>
> Eliot
>
> [*] I'm reminded of another example of this sort of thing.  Many years
> ago, in the middle of winter, my parents' house was broken into.  They
> called the police and the police quickly arrested a person who lived
> nearby, having tracked him by his footprints in a snowy field that
> links many of the houses.  Had he simply walked on the clear sidewalk,
> he probably wouldn't have been caught.  Not all criminals are super
> geniuses.[1]
>
> [**] Elsewhere it has been asserted that *there is no* ability to
> create a functioning escrow approach that scales.  I think that's very
> likely true, but if it turns out not to be, we should revisit 1984 at
> that time, as to whether or not it remains appropriate as a BCP.
>
> [1] https://www.flickr.com/photos/ladystephanie/10227056515
Sorry about that.

signature.asc
Description: OpenPGP digital signature