Re: How do I know if its encrypted?

At 10:08 PM 1/11/95 -0800, Eric Hughes wrote:
> ... Seems to me that a quite reasonable condition of use of a remailer is
>that what is passed isn't human readable.

Perhaps I missed this, but why?  If someone is going to plant kiddie porn or 
whatever on you, does it really matter if they encrypt it first or not?

If the purpose is simply to generate additional encrypted traffic to obscure 
stuff that needs encryption, that goal might be better served by simply 
encrypting files for their recipient as they pass through, in those cases 
where a public key is available and the msg isn't already encrypted.

I forget the name of the cypher (Vigere, perhaps--the one that uses a series 
of Caesar-like cyphers keyed by a password), but you could just run it 
through that with a password of abcdefghijklmnopqrstuvwxyz and you'd flatten 
out the distribution enough to get it by casual inspection.

    --Paul J. Ste. Marie
      pstemari@well.sf.ca.us, pstemari@erinet.com

At 10:08 PM 1/11/95 -0800, Eric Hughes wrote:
   > ... Seems to me that a quite reasonable condition of use of a remailer is
   >that what is passed isn't human readable.

   From: pstemari@erinet.com (Paul J. Ste. Marie)

   Perhaps I missed this, but why?  If someone is going to plant
   kiddie porn or whatever on you, does it really matter if they
   encrypt it first or not?

If you can't read it, it's not kiddie-porn *for you*, although it
might be for someone with the key.

Encryption fragments meaning subjectively.  A magazine, for example,
has a fixed center of meaning for all who can read the language.  A
magazine looks the same to all who look at it.  An encrypted file
looks different to those who have the key from those who do not.

Encrypted data is fundamentally different from paper-and-ink data in
this way.  The metaphor of "planting it on somebody" does not apply to
data that the "somebody" can't read.

   I forget the name of the cypher (Vigere, perhaps--the one that uses
   a series of Caesar-like cyphers keyed by a password), but you could
   just run it through that with a password of
   abcdefghijklmnopqrstuvwxyz and you'd flatten out the distribution
   enough to get it by casual inspection.

Fine.  It think that would suffice.  If you can't easily read it, you
can't be expected to have read it.  The operator of a data service has
_zero_ motivation to cryptanalyze something.  If they happen to apply
a viewer to the file (for whatever reason), they don't _want_ to see
what's inside.

Eric

In article <9501130137.AA03281@eri.erinet.com>,
pstemari@erinet.com (Paul J. Ste. Marie) wrote:
> At 10:08 PM 1/11/95 -0800, Eric Hughes wrote:
> > ... Seems to me that a quite reasonable condition of use of a remailer is
> >that what is passed isn't human readable.
> 
> Perhaps I missed this, but why?  If someone is going to plant kiddie porn or 
> whatever on you, does it really matter if they encrypt it first or not?
> 

The goal is to convince the two groups of concerned parties that the
remailer operators don't know the contents of what's passing through
their remailers:

         (1) the people who use the remailer, who get a measure of 
             comfort from knowing their communication is secure

         (2) legal groups etc. who may try to hold the remailer 
             liable in some way for what passes through their remailer.

A large percentage of material that passes through remailers might
be offensive to SOMEONE---if even just because an unpopular opinion
is expressed.  

The remailers are operated by people who want to promote information
flow, not restrict it.  They provide an important service that is of
critical importance to some people and groups who use the net.  They
shouldn't be held accountable for the few who abuse the remailers,
and encryption helps prevent that from happening.

From: lce@wwa.com (Larry E)

   The goal is to convince the two groups of concerned parties[, in
   short, users & lawyers,] that the remailer operators don't know the
   contents of what's passing through their remailers:

This is exactly right.  With a sealed box which you can't look in at
all, this is easy.  Providing an assurance on a general purpose
computer is more difficult.  And yes, it _is_ always possible to
simulate a filter that's not a filter, blah, blah, blah.  We are in
the realm of social interactions here, not in the realm of technology.

   The remailers are operated by people who want to promote information
   flow, not restrict it.  They provide an important service that is of
   critical importance to some people and groups who use the net.  They
   shouldn't be held accountable for the few who abuse the remailers,
   and encryption helps prevent that from happening.

I agree with this argument.  It is the germ of discourse about the
public policy of remailers and anonymity generally.  I want to point
out the rhetorical content of this statement, though, more than my
agreement with it.

The cypherpunks list is filled with paranoid nay-sayers who can't
distinguish their own paranoia from a legitimate technological
failing.  I feel a dire need for a positive rhetoric of cryptography.
I want to be 'for' something and to know what it's good for rather
than to be against everything that doesn't meet my personal desires.

How many times have I seen particular solution whose response is "But
I want more, and this won't work for that"!  The most self-deceptive
say "It can't be done", the slightly more honest say "You can't do
it", and none say "I will not do it".  So now all you people who think
that remailers don't work, don't run one.  Good, I see most of you are
already complying with this directive.

Even the simplest remailer has utility.  If there were no utility,
then nobody would use them (duh).  It is not only foolishness and
idiocy but also mendaciousness to say that "remailers just don't
work".  It is constructive to say, however, that "the current
remailers don't work against the following opponent", but this is not
usually the case.  Rather, the speaker's paranoia silently projects
their own requirements onto a technical discussion, leaving only
confusion.

Look at the recent conversation over postage for remailers.
Paraphrasing: "Credit cards won't work because they're not anonymous".
My response: "Bullshit".  Using a credit card as a means of payment
does put constraints on usage, but it doesn't prevent usage (duh
redux).  What credit card payment does do is to require more effort in
order to link email transactions.  This is an unalloyed good, but pure
silver instead of gold.  There are better ways, one of them First
Virtual, which at the least has counterparty anonymity; another, blind
sigs (as yet unusable for payments).

The implicit assumption here is that "If I can't use it to smash the
state, it's worthless".  Well, thank you very much for constraining my
ability for privacy with your political agenda.  And I have a hint for
all the state-smashing wannabe-businesspeople out there: the ones who
have a business (less secure) now will eat your lunch for the business
(more secure) later.

To be dry and academic about this, I'd say that the problem was an
insufficiency in threat modelling.  But that just doesn't quite mean
the same thing, n'est ce pas?

Eric