Make Your Own Money

I've been thinking about the internet, societies, trade, money, and ecash on
and off for some time. After chancing across a few urls that had some
interesting ideas about currency and de-centralised systems, I pieced
together a few ideas I thought might be worth sharing on how to implement
ecash.

Note that none of this stuff is my field of speciality, so I should
definitely _not_ be considered an authority on this stuff. There are bound
to be huge flaws that anyone with half a clue could spot. This should be
considered just the beginning of an idea, and perhaps a bad one at that.

In the article on "demurrage" currencies, xxx advocates a "stamp script"
currency backed by something of real value. The only problems I had with
this currency was; it was a "centralised", "paper money", currency. I'm a
net-head who has a distrust of anything centraly administered. I want my
real-value-backed demurrage currency to be de-centralised and network
transferable.

I'd been thinking that pgp signatures could be useful for an ecash for some
time. I hadn't thought about backing it with something of real value untill
reading the demurrage article. If I was issuing my own currency, what would
I back it with? After working all day, helping two more friends with their
computer systems, going home and realising that in 2 years I still hadn't
put a gate in the back fence, it became clear; my time is my most valuable
asset. What better to back a currency with? I'll give you an hour for your...

How do we depreciate our currency? Accountants would probably specify an
annual % depreciation. Being a physics kind of guy, I prefer to use a
"half-life", but one can be calculated from the other so they are really the
same thing. Depeciation of the currency can be easily justified in real
terms... "I promised an hour of my time, but that was two
years ago. Two years of my life are gone since then, so I have less time
left to give. After 16 years I could even be dead, making it useless."

Little is said in the "demurrage" article about what kind of depriciation
rates should be used. An ecash should be flexible.

So, we 

Attributes of Money
	medium of exchange
	measure of value
	store of value... undesireable?
	
Requirements of money
	robust		must survive many transactions
	divisable	large denominations can be split
	transferable	can be exchanged
	unforgeable	cannot be faked or copied

Any form of ecash is simply information. For ecash to work, information must
meet the requirements of money. 

Information is relatively robust; it's as good as the storage medium you
use. Transferability is easy, information flow is not much of a problem.
Divisiblity is also easy with information, but can actualy be a problem with
physical currency; how do you break a $20. Forged money can be divided into
two catagories; faked, and copied. At first glance unfakeablitiy appears
difficult, but the magic of digital signatures mean that this is actually
not a problem. Uncopyable; bzzzt, not possible, and is even a problem for
physical money. Despite many attempts by commercial industry to implement
copy-protection, the basic laws of information mean that if it can be read,
it can be copied. If it can't be read, it can't be used.

So how do we get around this serious limitation of information as a way of
implementing money? We need to minimise the risks of money being copied. Who
can copy information? Anyone who can read it. Who can spend copied money?
We can limit it by making the money like a non-negitiable cheque... that way
the only person who can benefit by copying money is the person it is signed
to.

Any transaction involving money requires a level of trust. We trust the
issuer of the money to back their currency. We trust that the money has not
been forged. This means we need to trust the issuer and each and every
person who had the opertunity to forge that money. This includes every
person who used that money in a transaction. How do we do that? First we
need to know everybody who used that money. This means money needs an "audit
trail".

For some time now the technology of digital signatures has identified that
one of the largest risks is not "cracking" the crypto, but using fake
identities. To reduce this risk the concept of a "web of trust" is used,
where digital signatures are applied to public keys to certify that "this
public key does indeed belong to whom it claims". Once you know you have the
public key of one person who you trust, you can probably trust any keys
signed by them, and probably also any keys signed by them, and so on. This
gives the "web of trust", helping ensure that you can trust the identity of
a public key even if you don't know the individual.


ecash=(halflife|ecash),date,serial,length,comment,tokey,fromsig

halflife=duration
date=time
serial=integer
length=integer
comment=string
tokey=gnupg public key of recipient
fromsig=signature of transferer

value= length/2^((now - issue_date)/halflife)