Re: King Kong Does e$

[Hal <hfinney@shell.portal.com>]

Thanks to Bob Hettinga for typing that long message about the Microsoft
"ecash" scheme.  That is some nimble note-taking.  I have a few comments
on the scheme as Bob presented it, and as it compares to Digicash.  I
will follow up with some commentary about the politics involved.

>         Customer chooses secret s, serial number n, n=h(s) and amount.
>         Customer pays bank (by any method) the amount of the desired banknote
>         bank signs serial number and amount
>         Transaction may be anonymous (physical cash purchase at bank) pick
>         serial number, bank uses private signature to sign the serial number.
>[...] 
> How is it used for payment
>         Payer simply gives bank note to payee by handing over the secret
>         payee immediately exchanges the banknote for new ones anonymously, if
>         desired
>         Old banknote is now invalid, since bank has its secret; hence payer
>         can't respend it
>         If the bank note is already invalid, bank rejcts exchange and informs
>         payee.
>         "In effect, these are disposable banknotes"

The withdrawal protocol has some similarities to online Digicash ecash.
In that system, you choose a random number s, calculate a one way
function h(s), and get that signed by the bank.  Unlike in the Microsoft
scheme, blinding is used for the signature.  I imagine Microsoft avoids
blinding because of the patent situation, and possibly due to legal
concerns about anonymity (more on this below).  With Digicash, the coin
is then the pair s, SIGN(h(s)).  This is then given to the shop as
payment.  It can check the bank's signature, but that is not enough;
being an online scheme, it must also turn the coin in at the bank to
prevent double spending.  The bank checks the signature and that the coin
is well formed to accept it.

The Microsoft scheme is like an unblinded version of this.  The bank
simply signs h(s) and gives that to the customer.  This allows a
simplification in the spending.  Instead of passing s, SIGN(h(s)), it is
enough just to pass s.  The payee gives this to the bank (since this is
an online system), which given s can calculate h(s) and check this
against a list of all valid coins.  It knows the valid coin numbers
because it saw them when it signed them (unlike with Digicash).  So there
is a slight space savings in the spending protocol.

It is also not necessary for the messages to and from the bank to be
encrypted during the withdrawal protocol; neither knowing h(s) nor
SIGN(h(s)) will allow an attacker to spend the coin, since he doesn't
know s.  A similar thing is true of DigiCash, though, where the blinded
pre-signed or signed coins are useless to an attacker since he doesn't
know the blinding factor.

The big problem then with the Microsoft system is that it is not
anonymous.  As a result, it is technically not electronic cash, at
least as the word is used in the literature.  However we are seeing so
many proposals like this, all of them wanting to capitalize on this
magic word "cash", that I suppose the definition has to be considered
to be shifting.  In the new usage, virtually any payment system can be
called cash if there is some way that users can be anonymous in using
it.  And since by allowing anonymous accounts virtually any payment
system can do this, the word is becoming meaningless.

The problem I see in practice with using their cash anonymously is how to
buy it.  If I have an account with the Bank of Microsoft, and I withdraw
some "mcash", deducting it from my account balance, that mcash will be
linkable to my account when I spend it.  In order to be anonymous I have
to buy the cash anonymously.  I can walk into the local bank with a
floppy and some dollar bills, but that is not practical in general.  I
could use mcash to buy some more mcash, but even if the second
transaction is anonymous, the bank knows that I was the one who withdrew
the first set of mcash, so it can link me to the second set when it is
spent.

The only good solution I can see is to use Digicash ecash to anonymously
buy Microsoft mcash, but I doubt that that is what they had in mind!

Frankly, what I see in this message is another example of something which
is starting to become common: marketing to cypherpunks.  In a way it is a
very positive sign, that our views and concerns are becoming so well
known and widespread that companies like Microsoft and Netscape are doing
their best to keep on the good side of people like us, who are concerned
about strong privacy and security.  In some ways our attitudes are
becoming dominant on the net, thanks to the many excellent writers here,
as well as magazines like Wired, and groups like the EFF and other
interest groups.

But this influence is making us a target of companies who know that
gaining our approval, or at least avoiding our criticism, is important
for success on the net.  In many cases, such as the recent flap over
Netscape's attitudes towards key escrow, I detect a whiff of two
sidedness, in which one attitude is presented for the benefit of
government and law enforcement interests, while another posture, more
acceptable to cypherpunks, is adopted on the net.  With Microsoft, they
use the magic word "cash" a great deal, in my view hoping that we will
line up in favor of the idea.  But as I have explained it is not really
anonymous, no more so than any other payment system.  And it is not at
all clear that the kinds of anonymous accounts that would be necessary
to really make it anonymous will be allowed.  In that case, Microsoft
can just shrug and say, "well, we tried."  They get the best of both
worlds.  They make the government happy by providing a traceable
payment system, while they look good on the net by pushing "electronic
cash".

I don't have any proof that this is exactly what is going on.  But it is
possible, and I think we have to be skeptical and at least open to the
possibility that this kind of manipulation is occuring, no matter how
many assurances we get from the companies involved that they are really
on our side.  Finance is a high stakes business and there is a lot of
government regulation involved.  Where our interests and the government's
diverge, we need to watch closely to see whether the companies' actions
match their words.  This kind of marketing is going to continue to increase,
I expect.

Hal Finney