From: rah@shipwright.com (Robert Hettinga)
Date: Sat, 2 Dec 1995 03:16:47 +0800
To: semper.fi@solutions.apple.com
Subject: GAK, Netscape, CyberDog, and you.
Message-ID: <v02120d0dace4adfb1c6c@[199.0.65.105]>
MIME-Version: 1.0
Content-Type: text/plain


Hi Semper Fi-ers!

I've been lurking on this list for a long time, and I haven't said much
here primarily because I'm not a developer. I mean, I've hired developers,
and sold the software I've had them build for me. Not very well, I might
add, which is why I try do something else (talk and write) for a living
now. Frankly, I couldn't code my way out of a paper bag, which, of course,
is why I've not said much here. :-).

Anyway, something has come up which I think makes the world very
interesting, and I think you should know about it.

Netscape has come out in favor of what the government calls "key escrow",
but which most "cypher"-punks (and I seem to have become one over the last
year and a half) call GAK, for Government Access to Keys.

Here's the URL: http://www.cnet.com/Central/News/govt.html . This is a news
report of a speech that Jim Clark gave at a trade show in Boston this week.

Normally, as serious as this is in the rest of the world, that wouldn't
really rate a mention here, except for the fact that this presents a very
interesting set of opportunities for the *independent* Mac developer
community in the form of CyberDog, Apple's set of OpenDoc parts for the
internet. I'll get to that in a moment, but first, let me give you some
background on the problem.


Netscape's Jim Clark has quite a tightrope act going. He has a stock price
which gives Netscape a P/E ratio of something like 7,000 (the S&P average
right now is about 14), which means he really ought to get some revenue in
the door, or his investors are going to have his hide. The next thing is,
the government is a *really* big customer, and *they* want GAK, in case you
haven't noticed. ;-). Couple this with the fact that Netscape is pretty
much replicable by concentrated developer effort because its underlying
technology is an open standard, *and* the fact that any significant attempt
by them to create any real proprietary standards on the internet is
practically impossible in the long run, unless they get a whole lot of very
big customers in a hurry.

Now, it looks like that's what happened already, because a lot of very
large companies have signed on to using Netscape servers, but you can also
see how really weak Netscape's position is when Uncle Sam knocks on their
door and asks for a key escrow "bug" in every "secure" Netscape web
session.

This isn't the first time that Netscape has had problems with financial
cryptography and the law. Netscape made the papers recently because the
government-mandated SSL key-size in their export version was too small. (We
won't talk about the cognitive dissonance between the words "export" and
"internet" just yet...)  The "export" version of  Netscape Navigator got
cracked by a French grad student in a university computer lab, who broke it
just by running a bunch of background processes on his Sun workstations for
a couple days over a weekend. The ITARs, the arms export laws covering this
(cryptography is legally a munition, more cognitive dissonance), are a now
problem for everybody who wants to do business securely and safely on the
internet.

More to the point, since digital bearer certificate technology, the most
economical method for doing business on the net, is entirely based on
digital signatures, which in turn are entirely based on very strong public
key cryptography, all business on the internet will eventually require very
strong cryptography. In other words, if you don't have unbreakable key
sizes, you can't issue digital certificates to pay for things with. It
would be like having paper money which is easy to forge.

Lately, I like to say that "Digital Commerce *is* Financial Cryptography",
and it's true. First Virtual has the only truly "out of band" internet
transaction settlement mechanism, it's very well-designed, and very robust,
but you have to remember that they did it *because* of the ITARs, and the
complications for financial cryptography those regulations cause, no matter
how much First Virtual itself likes to "dis" financial cryptography as a
concept. Unfortunately, the transaction costs on such out-of-band methods
are always going to be higher because of the inefficiencies of not
instantly settling transactions, like you can with something like digital
cash. When you buy something in the store for cash, you initiate, settle
and clear the trade all at once, right there at the cash register, without
an audit trail to maintain. Frankly, *any* credit card trade is really an
"out-of-band" method, because it has to go off of the net to clear and
settle, with audit trails, and the overhead of their offsetting
book-entries.

Fortunately, as governments start to figure this out, they will have to
stand back and let the economic train go by, so to speak.


So, having said all that, let's talk about OpenDoc and CyberDog.

I hear that with CyberDog, building a secure Netscape-compatible browser in
OpenDoc is now pretty simple. That adding economically useful -- and
interchangable -- functionality like strong encryption, digital cash
protocols and most of the tools of digital commerce is also pretty
straightforward. That is, even though they have to be developed separately,
because Apple justifiably doesn't want to have the ITARs limiting its
export market for both OpenDoc and CyberDog. Again that cognitive
dissonance between "export" and "internet".

In addition to building utterly secure browser-server links, I see some
interesting OpenDoc digital commerce applications coming out of a
Netscape-compatible browser. My favorite "flash" on this was imagining the
ability to drag Digicash ecash dollar bill icons out of a wallet and
dropping them on a cash-register icon in a web page to pay for a
transaction. That's certainly doable with OpenDoc and CyberDog.

In addition, the IETF has just promulgated a secure link-level encryption
standard called IPSEC, which allows for encrypted links between any two
machines on the net, with any cryptographic method you want. That looks
like a great Open Transport Streams Module project to me. And the wierd
thing is, once IPSEC is out there, GAK in Netscape becomes moot, anyway.
Netscape can only shoot itself in the foot here, and it looks like they
already have, with a premature announcement of GAK.


One final thing about Netscape. It's not their fault. There is no evil
man-behind-the curtain in all of this. The government's doing what it
thinks necessary for the preservation of order, and thinks it needs GAK to
do that. Netscape is doing the best it can for its (newly rich)
stockholders, and thinks it needs to comply with Uncle Sam, which it
probably in fact has to do, or burst it's stock market price sooner, rather
than later.

Also, there is no reason why Netscape can't make a Navigator OpenDoc part,
and I expect that they are planning to. I bet that Apple would really like
that, and as a matter of fact, is probably courting Netscape to do exactly
that.

But I also know that by using OpenDoc and CyberDog, that the user can
create a very easy to use, flexible, extensible, and *powerful* internet
environment. Being there first with a Netscape-compatible browser,
especially one that doesn't have GAK built into it, will be worth a whole
lot in the marketplace for any enterprizing independent OpenDoc parts
developer.


I'm going to be giving a talk on Wednesday at noon at Apple's Town Hall, (4
Infinite Loop) in Cupertino about what I'm calling "geodesic" markets, that
is,  financial cryptography on the internet.  The URL for a web page with
all the details is

http://thumper.vmeng.com/pub/rah/talk.html .

In addition, if you want more stuff on financial cryptography, you might
want to try my e$ home page, which is in my .sig, below.

Finally, I've asked an *actual* financial cryptographer, Eric Hughes, who
many of you may know from the cypherpunks mailgroup, and from the Clipper
fight, to show up at the Cupertino talk and bail me out on the real hard
questions.


Well, that's about it. Oh. The talk is titled "Financial Cryptography for
Dogs". I think you can see why, in light of the events of the last week,
it's quite appropriate.

Cheers,
Bob Hettinga



-----------------
Robert Hettinga (rah@shipwright.com)
e$, 44 Farquhar Street, Boston, MA 02131 USA (617) 958-3971
"Reality is not optional." --Thomas Sowell
The NEW(!) e$ Home Page: http://thumper.vmeng.com/pub/rah/
>>>>Phree Phil: Email: zldf@clark.net  http://www.netresponse.com/zldf <<<<<