What (The Hell) Is A Smart Contract?

Another good meeting of the BitcoinBlockchain Leadership Forum today, with the focus on practical use-cases for distributed ledgers and grasping at the nebulous concept of 'smart contracts' (links are to my own recent posts on these topics). 

In particular, we saw good, productive tension between Bitcoin blockchain purists who are intent on coding pretty much every element of a transaction into the blockchain; and those who see distributed ledgers as (also) playing a more limited role as just one layer or component in a broader array of gadgetry involved in any contractual scenario.

In my view, both approaches are valid but which 'wins' will depend on the use-case. And the development of the Internet demonstrates the technology will be used in ways no one intended anyway.

So, for my money, the definition of a 'smart contract' needs to be very broad, and I've suggested:

"an agreement performed via any number of applications, devices, networks and messages, which may involve entries in a distributed ledger."

This definition flows partly from a great discussion I had with Alex Amsel of Bitshake recently. I made the point that distributed ledgers seem most useful where a specific item is somehow dealt with or used very frequently and by many people or entities. Alex added a third condition: the participants are running different proprietary software, operating systems and/or devices - in other words they have an expensive interoperability challenge.

So a 'smart contract' might just be written in Word format, or html, and not embedded in a distributed ledger at all. But the subject matter of the contract - the rights to play a song, or rent a shipping container or space on a truck - might be 'hashed' into the ledger, and users' machines could interact using that hash, triggering instructions to pay the contractual amount to a certain account. Multi-factor authentication as one step in the contractual process (e.g. identify checks for anti-money laundering) is another example.

At the forum, there was mention of locating, booking and paying for a car space as another example. This was dismissed by lots of people who said you can already do this without a distributed ledger - the parking space is already entered in the systems of the council's chosen payment service provider. But that means I need to know which municipality I'm in to find the right payment app, download it and register a payment method before paying (I changed cars recently, so I have to re-do all that). And that inaccessibility is partly a function of having to cover the cost of expensive proprietary systems. But if parking spaces were 'hashed' in an openly accessible public ledger, couldn't our smartphones find and pay for them using that code and our own chosen payment method?

Anyhow, the point is not that we necessarily need distributed ledgers to pay for parking or any other specific use-case, but that once people begin using distributed ledgers more widely, tons of other apparently trivial uses become feasible and worthwhile. Conversely, a comparatively trivial but widely shared use-case might unleash more widespread adoption, as happened with text messaging (I'm not suggesting that parking will do it, by the way).

Of course, Bitcoin users will be screaming at their screens by now, if they've got this far. They'll be shouting that Bitcoin has already unleashed distributed ledgers. 

They're probably right.