[Cryptography] Decentralized, global, irreversible, transaction database...

[Jerry Leichter]

On Dec 21, 2013, at 1:52 AM, ianG wrote:
> The smart contracts idea was pioneered by Nick Szabo in the 1990s, and was not credited in the original paper.  It's now credited by Mike Hearn in a wiki post:
> 
> https://en.bitcoin.it/wiki/Smart_Property
Clever but remarkably naive about how the business world actually works.

Some personal assets of high value - real estate and cars are the common example - are only owned through titles.  A title is really just a published document, accessible through (in the US; there are analogous systems elsewhere but I know nothing about how they are implemented) a state- (cars) or locally-(real-estate) provided repository, that witnesses to everyone who wishes to check who actually owns the property in question.  To transfer ownership of such property, you need to "transfer the title":  Publish to the repository, using approved procedures, that the title has now been transferred to someone new.

For some other kinds of assets there are non-governmental systems that provide the same role.  Artwork is worth very little unless you can "establish provenance":  Show a chain of ownership going back to the original artist.  Private organizations (art dealers, auction houses) authenticate these chains of transfer.

What keeps a business from going to five different banks and pledging the same factory equipment as collateral on five different loans?  Every state in the US provides for "UCC filings" - a formal way for a lender to declare that some piece of property has been pledged as collateral.  (UCC is the "Universal Commercial Code" - a set of standardized laws for commercial transactions that in one form or another have been accepted by most of the states.  The ones that haven't accepted the UCC have very similar mechanisms in place.  Without this uniformity, inter-state business would be very cumbersome.  There are, in fact, even some international standards - e.g., everyone in the shipping business world-wide agrees on what "FOB" means.)  Any lender will "check the UCC's" before proceeding with a loan on business equipment.  (This isn't needed for real estate or cars because the loans are recorded on the title instead.)

So what Szabo proposed isn't new; it's implementing an idea that goes back almost 1000 years (see the Domesday Book) "with a computer" "on the Internet" (like all those horrible business process patents).  The old mechanisms relied on centralized, trusted third parties (usually local governments) to maintain the list of transactions - which were always public.  The proposal is to remove the trust and the third parties by using cryptography.

FWIW, it's not clear how good an idea this is.  Right now, the groups maintaining the lists are closely tied to the ones enforcing the contracts implied by those lists.  If you continue to rely on the courts to enforce the contracts, it's not clear you gain much from cryptography.  On the other hand, if you move in the direction of "self help" - as in proposals that cars be automatically disabled if you don't pay your loan ... well, we've already seen moves in that direction, and they've proved to be highly contentious:  Easily abused, providing little recourse in case of error.  A car that your bank can shut down remotely when it thinks you haven't paid is a car with DRM.  Do you *really* want that?
                                                        -- Jerry