thoughts on digital cash

[Marc Horowitz <marc@MIT.EDU>]

I've thought a bit about digital cash since the thread started here,
and I have a bunch of thoughts.  I hope I'll be coherent.

** For production use, you want a real, live bank or other entity to
issue the "money".  Assuming the digital cash is backed by gold, or
securities, or something else tangible, you don't want the bank to go
away with all your money.  Or, if you read in the newspaper "Feds
seize First National Bank", you want to know if they've just seized
*your* bank, and therefore will be able to watch your deposits and
withdrawals.  If you don't know who your bank really is, you won't
know.

** All of the above applies to your escrow agent (who may be your
bank) for similar reasons.  Public key systems provide
non-repudiability, but knowing provably that a pseudonym just stole
money or goods from you doesn't help you get it back.  Sure, it
tarnishes their reputation, but if they steal a whole lot of money all
at once, they won't care.

** Anonymity in banking is a well-known concept.  If I transfer money
from my numbered swiss bank account to your numbered swiss bank
account, nobody knows anything, except who the bank is.  And I don't
see Switzerland giving up their enviable position in the international
banking community willingly.  All we need to do is convince a swiss
bank to do some digital cash banking.  Easier said than done.

** Legality of starting our own anonymous electronic bank: what do the
laws say, anyway?  As long as I pay taxes on any profit, I can form a
sole proprietorship to freely buy and sell purple widgets with very
little government interaction.  Would it still be legal if I got a
whole lot of people to accept purple widgets instead of cash?  Would
it be legal if we called it "digital purple widgets" instead of
"digital cash"?


I'd really like to see some sort of digital cash system set up.  My
feeling is that there would be a stronger demand than most people
think for such a thing, even if it weren't anonymous and untraceable
and all that.  But if it has those features, too, and we do it first,
it will get used.  And de facto standards are hard to get rid of.

		Marc