Underestimate blockchain at your peril [LWN.net]

Underestimate blockchain at your peril

Posted Jan 3, 2018 14:51 UTC (Wed) by smurf (subscriber, #17840)
In reply to: Underestimate blockchain at your peril by anselm
Parent article: Welcome to 2018

This is an advantage, in that the fact that you can't change smart contracts will *finally* teach people to do due diligence, correctness proofs, best practices etc. when coding them, or else they'll have to face the legal consequences and liabilities of shoddy work – just like in any other "real" enterprise where your builder or electrician's asses can be sued off when they make preventable mistakes.

to post comments

Underestimate blockchain at your peril

Posted Jan 3, 2018 15:44 UTC (Wed) by anselm (subscriber, #2796) [Link]

Possibly. But how likely is that to actually happen?

Anyway, in real life, it is the intent behind a contract which counts in the end, not the language of the contract. “Smart contracts”, though, are all about the language of the contract, and implementation bugs make it likely that the “smart contract” does not accurately represent the intent of the contracting parties but can't really be fixed. Actual real-life contracts, OTOH, usually contain provisions that arrange for unenforceable bits of the contract to be replaced by such enforceable bits that are closest to the intent of the parties (where what is enforceable and what isn't may only come out after the fact, e.g., because the regulatory framework changes). How is that dealt with in “smart contracts”?

(It also turns out that “smart contracts” can in fact – surprise! – be changed or voided retroactively if the people in charge of the smart-contract platform would otherwise lose money or dislike the contract for other reasons, as in the “The DAO” incident on Ethereum. Not a good sign.)

Underestimate blockchain at your peril

Posted Jan 3, 2018 21:57 UTC (Wed) by roc (subscriber, #30627) [Link]

Builders and electricians make mistakes all the time. Those mistakes can be fixed, and frequently are, and it's usually no big deal.

Arguing that all we need to start developing perfect software is the right motivation seems ... optimistic.

Even if it was true, there's a cost consideration. Historically, applying our best tools for software verification to every piece of software has been prohibitively costly. If writing smart contracts is more expensive than hiring a lawyer to write a contract the old-fashioned way, people won't do it.

Underestimate blockchain at your peril

Posted Jan 4, 2018 20:33 UTC (Thu) by nix (subscriber, #2304) [Link]

Meanwhile, in the real world, the predominant language for writing smart contracts was modelled on JavaScript, and the docs for it (insofar as there were any docs) said how easy it was to write code for it... and the first (and second, and third) major uses of smart contracts all imploded with more or less all the money stolen by hostile attackers and/or people just playing due to bugs that could not be fixed.

Underestimate blockchain at your peril

Posted Jan 6, 2018 2:03 UTC (Sat) by elvis_ (guest, #63935) [Link] (1 responses)

Excellent!!! So a standard contract will cost twice as much because of all the unneeded work involved to make it perfect. A terrible outcome for 95% of people. Have you ever been in the real world?

Underestimate blockchain at your peril

Posted Jan 6, 2018 13:15 UTC (Sat) by smurf (subscriber, #17840) [Link]

The "real world" is in fact a very good example. Real-world processes contains checks to make sure that nobody empties your bank account or sells your home without your consent, and procedures to undo the damage when the checks fail, or when the process is deemed invalid or illegal after the fact. Code in the blockchain needs to be able to do the same thing. That may not always be possible (you can't take Bitcoin/Ether from me without my consent) but a contract can do something equivalent (hold some coins in escrow, create the digital equivalent of title insurance, create a token which you *can* take away from somebody when legally mandated, …).

Which "unneeded work" would there be in a standard contract? Checking whether the parties involved have enough assets / sufficient authorization to execute it is hardly "unneeded". Besides, drafting and executing a contract in the real world costs way more than twice the coding and gas cost of a typical Ethereum blockchain contract.