Netscape as vehicle for cypherpunk agenda/the cypherpunk bully pulpit

I was thinking recently how the events of the past week or so
have turned me into a sort of a Netscape advocate. Granted, there are
bugs in Netscape, and there probably will be more bugs uncovered
(someone needs to write an exploit if they want themself & Ray to get
a T-shirt btw), but Netscape is interested in fixing problems and the
new 2.0 is doing encrypted email, probably with a really nice
interface (Haven't seen it yet, of course) and they are working to
make the 128-bit version downloadable. (The 128bit version is
available overseas already anyway, I hear.)

	The really big sticking point I see, however, is the
certification authorities. There is a single point of failure here and
that is at Verisign. This becomes a large problem I think if the en
rypted email that Netscape does requires personal x509 certificates (I
read that Versign is issuing those for $9/each.) This is a problem
because for one thing I don't think Versign will want to issue certs
to psudonyms, and Netscape may not talk encrypted email to
non-certified people. (I am not sure)

	The solution to this, of course, is to allow Navigator to
accept alternate certification hierarchies, so we can setup a
Cypherpunks cert agency or a c2.org cert agency, which -will- sign
nym's keys, etc. The question exists though, as to whether or not
Netscape will allow for alternate agencies in Navigator.

	I haven't seen any mention of this feature in 2.0, so if the
feature exists in 2.0, then great! Otherwise, unless Netscape is going
to allow for alternte cert agencies on a specific timescale, I think
we have to do something about it in order to force the issue.

	Along the same lines of what happened recently-- because of
the exposed hole and the pressure put on Netscape, management was
finally willing to let some of the code be available for public
review. If something happened to show how relying on a single point of
failure such as Verisign was bad and resulted in much press &
publicity, then perhaps Netscape management would be convinced to
allow for alternate cert hierarchies..

	Some sort of hack which demonstrates this would be great. I am
feeling uncreative and can't think of anything effective short of
stealing Verisign's private key, but that would be pretty damn tough.

-- 
sameer						Voice:   510-601-9777
Community ConneXion				FAX:	 510-601-9734
An Internet Privacy Provider			Dialin:  510-658-6376
http://www.c2.org (or login as "guest")			sameer@c2.org

I agree with Sameer's points here. Netscape carries with it the
potential of revitalizing the cypherpunks agenda, but also, in the
worst case, making it irrelevant. What happens, I think, depends on
what we do.

On the plus side, Netscape 2.0 will, without a doubt, be the first
usable mail tool to incorporate real encryption. It may accomplish,
almost overnight, the long-held goal of making a nontrivial fraction
of Internet email secure.

Another potentially big win is the Java language. It seems to me that
it will be quite plausible to code up real crypto applications in this
language. Once coded, these applications will run on every important
platform in the universe, and can be accessible by the click of a
mouse. Perl-RSA was a sign of what's possible in the non-C world. One
caveat is the slowdown from the interpreted code (roughly a factor of
25). The best way to look at this is as a challenge, to use clever
coding tricks and intelligent architectures, including caching. The
speed of Java implementations will inevitably improve - in fact, I
might just be doing my PhD thesis on memory management in Java.

One potential downside, as Sameer points out, is the X.509
certification hierarchy. This ancient beast has the potential to
defeat many of the cypherpunk aims, most especially the possiblity of
anonymous communication. However, that's not a foregone conclusion.
The most important thing to be doing right now is to _understand_
what's happening. Over the next couple of weeks, I'll be reading the
S/MIME and X.509 documentation, poring over reference code, and (of
course) playing with Netscape 2.0 myself. We're much more likely to
get our agenda implemented if we are armed with a good understanding.
There are lots of ways around X.509 - maybe we can work our way around
it, maybe we can adapt it to our needs, maybe we can come up with
something better and get it replaced. Whatever the case may be, we
should not give up hope.

Netscape is one manifestation of the ancient Chinese curse: may we
live in interesting times!

Raph

sameer wrote:
 
>         I haven't seen any mention of this feature in 2.0, so if the
> feature exists in 2.0, then great! Otherwise, unless Netscape is going
> to allow for alternte cert agencies on a specific timescale, I think
> we have to do something about it in order to force the issue.

Netscape has already annonced that for 2.0, the user will be able to accept (trust)
(or reject) any set of certificate authorities for signing of certificates.

I briefly looked for a copy of that announcement, but I could not find it.

PK
--
Philip L. Karlton                       karlton@netscape.com
Principal Curmudgeon                    http://www.netscape.com/people/karlton
Netscape Communications Corporation

In article <199509251741.KAA04656@infinity.c2.org>, sameer@c2.org (sameer) writes:
> 	The really big sticking point I see, however, is the
> certification authorities. There is a single point of failure here and
> that is at Verisign. This becomes a large problem I think if the en
> rypted email that Netscape does requires personal x509 certificates (I
> read that Versign is issuing those for $9/each.) This is a problem
> because for one thing I don't think Versign will want to issue certs
> to psudonyms, and Netscape may not talk encrypted email to
> non-certified people. (I am not sure)

  I believe that the identies of free certificates that verisign plans
to offer to netscape customers will not be checked in any way other than
to ensure that the name is unique for that CA.  You will have to ask
someone from Verisign to get a certain answer.

> 	The solution to this, of course, is to allow Navigator to
> accept alternate certification hierarchies, so we can setup a
> Cypherpunks cert agency or a c2.org cert agency, which -will- sign
> nym's keys, etc. The question exists though, as to whether or not
> Netscape will allow for alternate agencies in Navigator.

  I have stated here, and in other public forums, several times in the
past few months, that Netscape Navigator 2.0 will support user configurable
certificate authorities.  You will be able to specify that you do or
do not trust specific server certificates and certificate authorities.
The user will be able to incorporate new CA certificates into their
certificate database, and mark them as trusted for signing certs for
SSL, email, etc.

	--Jeff

-- 
Jeff Weinstein - Electronic Munitions Specialist
Netscape Communication Corporation
jsw@netscape.com - http://home.netscape.com/people/jsw
Any opinions expressed above are mine.