Re: [Endymail] spam versus cleartext

[Dave Crocker <[e-mail]>]

On 9/7/2014 7:54 AM, Pete Resnick wrote:
> On 9/7/14 11:09 AM, Dave Crocker wrote:
> Signatures, just like encryption, are part of cryptography. If you are
> doing cryptography (in the way we normally do so for e2e encryption),

At the level of "if you can get keys you can do either", sure.

Arguably the nature of the trust assessment issues is different for the
two, but that could get esoteric.


>> So you want to eliminate anonymous communications?  Anonymity has
>> historical importance for some kinds of communication.   
> 
> Pseudonymity (i.e., a signature that is not attached to a particular
> human identity) may be sufficient for most cases.

Might.  Might not.

We have little operational experience with some of these constructs in
the practical world.  I'm pushing back about all this because we need
much more cautious language about the efficacy and risks of these
approaches.

In effect, I suggest approaches be characterized as (potentially) useful
options, rather than likely or certain "solutions".  Given the way these
topics tend to be discussed, that distinction is fundamental.


> Doing so would still
> require a prior-to-real-communication step of me allowing that signature
> into my whitelist/contact list/whatever. For my personal email, I am
> perfectly willing to say,

You (and I and everyone else on this list) are not representative users.

Most of the human factors experience in this realm is that average users
don't appreciate the extra hassle and don't perform well with the
additional tasks.  So if you want these mechanisms to scale, they
require thinking very differently about end-user load.


>> My point is not that signing is bad or checking against address books is
>> bad, but that mandating such things constrains legitimate communication
>> in important ways.
> 
> Let's not miss the point that we are *currently* constraining legitimate
> communication in important ways, as my weekly hunt through my spam
> folder and my occasional out-of-band, "Why did my mail bounce?"
> complaint amply demonstrate. I choose my tradeoffs, I get the advantages
> and disadvantages of those tradeoffs.

You haven't heard me suggest maintaining the technical or operational
status quo and ignoring the problem, nevermind the amount of time I
spend in the world of m3aawg.org and more recently Levison's effort.

Again, my concern is ensuring adequate caution about unintended (as well
as intended) consequences.  The very consistent tendency of folk making
proposals in this space is to be quite cavalier about the human
communication downsides from imposing excessive constraints.

d/

-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net