Re: Bypassing the Digicash Patents

[Robert Hettinga <rah@shipwright.com>]

At 5:09 am -0400 on 4/30/97, Adam Back wrote:

> No offense Bob, but your pose takes some reading, too full of
                <vogueing> ^^^^ Freudian slip? </vogue>
> metaphors,

Consider yourself yet another victim of this philosophy major's penchant
for WFFy reason-by-analogizing (Wow. Almost as good as jya there...
<jya-as-Elvis>Thankyewverramuch</Elvis>).

> but I grok what you're saying, and the topic discussed here
> I find interesting.

Glad to oblige...

> How about this, rather than interface your ecash system with US
> dollars yourself through credit cards/ debit cards/ cheques / cash,
> just set up an entirely disconnected system.

Nah. I want to have real money backing it up. Any attempt to make money
less negotiable reduces its usefulness. Remember the Soviet Ruble? An
extreme example in the opposite direction, surely, but you get the idea.

> You may remember the digicash trial mint.  It was monopoly money,
> theoretically it was worthless.  However people were selling freebees
> for it (the odd T-shirt, cap etc), plus images, programs.  Also it was
> collectable in the sense that there was a limited mint.

Yup. Remember, it was Rich Lethin and I who set up ecm@ai.mit.edu, which
was a market where those digital cash certificate could be exchanged for
cash. Lucky Green sold the first ones, and Mark Grant(?) even put up a web
page to simplify things, using the list as a "tickertape" to announce
trades.

All of which actually proves my point. Because the market actually *did*
route around the lack of exchangeability. It's much better, of course, to
build exchangeability into a digital bearer certificate market from the
outset. Money's supposed to be negotiable, after all. :-).

What we have here is more a question of a business model rather than a
problem with cryptographic protocol.

> However this means you've got to trust the bank not to mint unlimited
> amounts of money for it's own use.

Right. That's why you have a separate trustee holding the reserve capital.
Again, it's using the right business model, and not necessarily
cryptography, which makes a market happen. Blind signatures and hash
collisions are necessary, but not sufficient, for the market to exist.
Anyway, in the first stages, I claim a trustee should be an actual, real,
um, "hoity-toity", bank. In the same way that SET and Cybercash and ATM
machines "blind" their transactions through the host bank onto a settlement
network to the customer's own bank, there can be sufficient blinding of the
transaction through the trustee so that the only thing the trustee sees is
a confirmation to pay and a settlement wire from the cash purchaser's bank.

Of course, at some point, the trustee can just hold other bearer
certificates instead of keeping the issuer's reserves in book-entry assets.
When there are other bearer certificates to hold, anyway...

How you issue those certificates mechanically is not nearly as important as
the fact that you *can* issue them uniquely. Ideas like hashcash and
micromint work real well for very small transactions, for example,
precisely because of the cost to generate the first one in the series,
which forces you to print a whole bunch of subsequent ones to pay for the
computational resources you've used. However, once again, um, no offense,
what cryptographic protocol you use to generate the certificate is the
functional equivalent of <analogy-warning> doodling, the process which
makes those complex graphic fills on paper currency which were designed to
moire up any attempt to photoengrave a certificate copy. </analogy> The
point is, you need cryptography for a digital bearer certificate market,
but it's not sufficient to create that market.

> But if you've got multiple banks then you've got to have an exchange
> mechanism.  The market could probably take care of this, setting
> exchange rates based on banks reputations.

Exactly. For instance, (hint, hint) if someone were to build
Eudora/Netscape/Quicken plugins for FSTC electronic checks, and a
plug-and-play deposit server for banks to receive them and convert them
into ACH transactions, who says you need the ACH system to settle the
checks anymore? All the different bank servers could just clear against
themselves on the net at some point, cutting their ACH fees out completely.
Someday.

> However it would be nicer to have something which required no trust
> and which had no posssibility of cheating rather than relying on
> reputation to sort them out.

Actually, I think there is no such thing as finance without reputation.
:-). I'd be very interested to see how you can prove otherwise...

> The problem with anonymous ecash to continue your metaphor is that
> .WK1 files also happen to be illegal or surrounded by huge amounts of
> banking regulations.  So even though the new system is better the
> negative forces acting against so far have succeeded in stifling it.

Nah. Reality is not optional. :-). Ask all the former 30xx COBOL-jocks out
there.

Remember, crime is orthogonal to technology. Bank robbery (these days,
anyway) is an artifact of automobile and firearm technology. The other uses
for automobiles are <metaphor-warning> cetacean in comparison to the krill
of "automobile crimes" like bank robbery </metaphor>. Firearms,
unfortunately, are in the same boat. Like nuclear power, we have bans on
guns because we can "afford" to have them. People don't hunt for food
anymore. I think that's why heavily populated countries have fewer guns.
You don't need them to eat, and even if you did, you'd starve, because all
the food was hunted out long ago...

BTW, it seems to me that hoplophobia, like innumeracy, is practically a
luxury. At least until <analogy-warning> the state shows up at your door in
black pajamas and asks you to put on some orange ones of your own for this
cool sleepover they're having a few miles out of town </analogy>.

> One of the negative forces also is user stagnation, people are used to
> cheques and credit cards, even if they are inefficient and prone to
> fraud.

I, for one, think the whole concept of "path dependency" is bunk, but let's
not clog the list with discussions of roman wheel ruts, QWERTY keyboards,
and the devine right of Windows...

Cheers,
Bob Hettinga

-----------------
Robert Hettinga (rah@shipwright.com), Philodox
e$, 44 Farquhar Street, Boston, MA 02131 USA
Lesley Stahl: "You mean *anyone* can set up a web site and compete
               with the New York Times?"
Andrew Kantor: "Yes."  Stahl:  "Isn't that dangerous?"
The e$ Home Page: http://www.shipwright.com/