- But Ethereum has also won fans with its promise to do much more than Bitcoin. In addition to the virtual currency, the software provides a way to create online markets and programmable transactions known as smart contracts.
The system is complicated enough that even people who know it well have trouble describing it in plain English. But one application in development would let farmers put their produce up for sale directly to consumers and take payment directly from consumers. There are already dozens of functioning applications built on Ethereum, enabling new ways to manage and pay for electricity, sports bets and even Ponzi schemes.
(title is the print title of the article, don't blame me)
mk, what do you think of this comment on Hacker News about Ethereum?
- Ethereum becomes useless as soon as it interacts with the real world. Its brave new world of digital contracts works fine for creating simple gambling games or 'provably fair' Ponzi schemes, but as soon as a piece of code on the blockchain has to rely on a real world piece of data, you lose all of the benefits of the system.
You want to write something related to (say) a share price, or property ownership? Then your code needs to trust an external system providing the data, and the users need to trust that other people will honour the 'facts' decided by your program on the blockchain.
You want to create even a simple system that opens a lock when someone pays money? (The much-hyped Slock.it startup selling 'smart' locks) The blockchain makes no promise that a physical lock will operate based on bits in the blockchain. You have to trust that someone isn't cheating you.
And why is this need for extra trust a problem? Because once you require trust, you are back into the real world where we already have real solutions to these Ethereum based problems, that rely on trust (e.g. Share dealing, property sales, etc). By moving onto the blockchain, you have gained nothing and only complicated matters.
I agree that the oracle problem makes many use cases for Ethereum over-hyped at present. Nevertheless, there are plenty of oracles that can be easily shaken out, such as the Fed's federal funds target rate, or the inflation rate as reported by the Bureau of Labor Statistics, or simply an asset price from an exchange or group of exchanges. There is a business for this, and http://www.oraclize.it/ is probably going to do well. There will be more. But that's not really the short- or mid-term promise of Ethereum, IMHO. Ether is programmable money, and within the context of the transfer of value via ETH, payment networks and mechanisms can be highly granular and complex because of the trustless and open nature of the ledger, and because the Ethereum scripting language is Turing-complete. As an example, Bitcoin Core developers are hoping to create payment channels for Bitcoin (called the Lightning Network), which will require many changes to the Bitcoin code, as well as at least one hardfork. Such payment channels can be added to Ethereum as is. In addition to being Turing-complete, Ethereum uses an account-based architecture for accounts balances rather than the UTXO (unspent transaction output) model that Bitcoin employs. The balance at an ether address is expressed in a constant number of bytes due to settled deposits and withdrawals, Bitcoin uses the sum of all unspent transaction outputs to determine the balance of an address, which can create problems for bundling lots of small payments on-chain as some UTXO transactions can be big. Also, Ethereum blocks are 14 seconds with little variance, whereas Bitcoin blocks are 10 minutes with high variance. In short, in just about every use case, Ethereum is better for transacting ETH than bitcoin is for transacting BTC. Also, you can simply use the Ethereum network as a ledger without intrinsic contract triggers that need oracles, like tracking the ownership of shares. EDIT: I went over to HN and dropped a reply. Jesus, those HN folk hate cryptocurrency. They have been powerful haters of Bitcoin all along, and it looks like to be the same case for Ethereum. It's very interesting, because you would think that a forum of programmers would be a generally receptive audience. There are a lot of misinformed and negative comments every time cryptocurrencies are discussed.
I agree with you that dApps will probably work fine. Augur seems very promising to me. What I don't yet fully understand is how digital contracts will be much more transparent than current contracts. As you said, contracts can be very granular and complex. Take, for example, a rental contract. How is replacing legalese with ether code going to make it easier to understand for everyone in the world who can't read code? Hacker News suffers from very reactionary comments. Most comments are just rebutting some of the arguments of the post or comment above it. The self-driving car comments there taught me that they're disproportionally sceptical when there are easy counter-arguments to make.
I don't think Ethereum will touch such contracts in the foreseeable future. More likely it will just be code for multi-party release of funds or changes in asset ownership, escrow accounts, and things like that. The community solar project using Ethereum in NY, is an example.Take, for example, a rental contract. How is replacing legalese with ether code going to make it easier to understand for everyone in the world who can't read code?
Any solution to this can be a problem, you can't just ignore everything after the ? Imagine a blog post with this URL: myblog.com/?post=152 Many websites depend on what follows after the ? to display different content, even hubski a few months ago used to do that. But I agree that a method to compare URL similarities is a good idea. I think this method should warn the user about previous posts or give them a list of similar URLs, ultimately the user will have to make the decision of posting that URL based on manual comparison with the list of similar URLs.
Isn't the point that the trust one is expected to rely upon is peer to peer, rather than hierarchical? You don't trust a MITM entity to mediate the data, you trust the cryptographically secure communication between the individual entities (buyer, seller, the lock itself) all mediated by open smart contracts enacted on the ethereum blockchain. Unless the objection is a really profound objection to the phenomenology of real word/digital interaction in which case that's a doubt that enters at the interface between the digital and ontological world.