Olympic: Frontier Pre-Release | Hacker News Reader

Olympic: Frontier Pre-Release

blog.ethereum.org

35 pointszmanian16 comments

I'm the release coordinator for Ethereum (although I'm best known for open hardware http://hexayurt.com I do software too from time to time.)

We're certainly getting close. As you may know, we have a "never set dates" approach to our release schedule. Pushing for a date threatens a burned out developer cutting corners which later turn into security issues, and we'd kick ourselves if we shipped with issues we could have avoided with a realistic schedule.

I say this because the pressure is always to say "two weeks" and, of course, the answer has to be when it is ready.

But, soon.

My own description of ethereum: it's a shared database that is shaped like the network, and you can store and run code in it, using a standard virtual machine.

Everybody describes it slightly differently.

Irregardless of what one may think about Ethereum as a platform, I have a lot of respect for Ethereum's hiring and retention strategy. Ethereum seems to have a knack attracting, on-boarding and deploying a deep roster of interesting individuals.

Thanks Vinay for weighing in.

Anyone want to give a summary of what Etherium is?

In one line: "Ethereum [is] the generalised blockchain for smart contract development" A summary probably won't cut it, you should take a glance at https://github.com/ethereum/wiki/wiki

An overengineered and useless cryptocurrency (Bitcoin-like currency). It was also pre-sold so they've made quite a bit of money for themselves (out of nothing) already.

Smart contracts are pointless since you need trust anyway to deal with real-world assets. Nobody would rather trade in a p2p system instead of a "real" exchange either.

Overengineering in Ethereum is rampant too, as it's trying -and failing- to definitively work around Bitcoin shortcomings by significantly increasing complexity without good reason. They're also reimplementing it on several programming languages, without code reuse from Bitcoin or any other related technology either.

Though I disagree with you, I upvoted you because your criticism is substantive and well informed.

Thanks. If only everyone did that on HN... Someone should teach people they're supposed to upvote stuff based on how useful it is, not on how much they personally agree with opinions expressed.

> Smart contracts are pointless since you need trust anyway to deal with real-world assets.

jgarzik (bitcoin core dev) disagrees with you, he tweeted recently that he's excited about their potential (he calls them "active addresses"). Also, Nick Szabo disagrees with you.

> Nobody would rather trade in a p2p system instead of "real" exchange either.

Lots of people would rather. There are many projects attempting to do decentralized/p2p exchange. For example, atomic cross-chain transactions have been on the horizon for a while now. And startups like tether.to, so that USD-tied assets can circulate in a decentralized manner (I won't mention ripple, but it is the same idea).

It is a natural evolution, the demand for an alternative to centralized exchanges comes from the same motivation to have an alternative to PayPal. Oh, and people also want a decentralized Silk Road (once they realize that its possible).

They haven't made money "out of nothing", the money is paying for development. They've had running code since before the Ether sale. That code has been used in projects like Counterparty and Eris. It's already of value.

Implementing a system in different languages ensures that, unlike Bitcoin, assumptions and bugs don't get baked in to a single codebase or development environment. It's a robustness measure and good software development practice (I'm disturbed by how few supposedly professional software engineers are familiar with this principle). Simply re-using Bitcoin code would work against this.

Smart contracts can reduce legal costs in trusted transactions. And they're not just for replacing existing legal contracts.

As for "real" exchanges vs. p2p, it depends what you're trading. Twenty years ago if we were discussing whether someone would one day sell a house on eBay I'm sure we'd be having a similar conversation.

The development process for the final release isn't finished yet so if it's true that they've "failed to work around Bitcoin's shortcomings" that's something you should go into more detail about so it can be addressed.

Think of legal contracts as executable scripts. The Etherium network is the infrastructure to make those digital contracts work in a decentralized and autonomous fashion.

So an example use case would be a script that receives transactions from anyone. Each transaction would include some coins and some output address. Then at specific intervals the script would dump all the received coins into the most popular of the received output address. It would effectively be sort of a rudimentary charity organization. The beauty of it is that enforcement is 100% guaranteed and automatic and there's also 100% transparency and furthermore the script has no owners or masters. It's completely autonomous.

The scripts are called DAO/C, Decentralized Autonomous Organizations/C..

The Ethereum project is conceived a crypto mega project which encompases a turing complete shared state machine(blockchain), content distribution and messaging.

The goal of Frontier and the Ethereum blockchain is provide a viable host ledger for non-terminating state machine.

The Bitcoin blockchain hosts state machines but they are guaranteed to terminate. An example of this is a unspent tx output hosted in a multisig smart contract.Spending the UTXO terminates the smart contract.

Ethereum is attempting to build a ledger system where the transactions update the state of smart contract without necessarily terminating it.

Applications:

If successful the Ethereum blockchain and other users of the Ethereum Virtual Machine technology should be able host arbitrary consensus and have updates to the consensus depend on arbitrary sets of cryptographic proofs like signed transactions, votes, zero knowledge proof of votes, attestations signed by pinned keys etc.

Challenges:

1. Gas

Ethereum proposes a gas value to compensate miners who host and compute updates to the ledger state and limit abuse of their computational resources by network users.

Optimizing the gas metric which sums computer,storage and network resources is NP Complete. It is an open question how closely one must approximate the optimum to have a healthy and usable system.

2. Centralization

Ethereum faces centralizing effects around proof of work, memory burden from hosting and updating smart contracts etc. This will likely lead of conservatives about what kinds of transaction the network will accept in practice and limit use cases.

3.Security

The surface area for attacks that trigger blockchain forks is vast and its difficult to full enumerate all the potential attacks and prove security

The much bigger news buried in this post is that they're suggesting the production system will likely launch in about 2 weeks.

Don't worry, the fanboys will spam HN then again.

Not exactly. Frontier will launch 2 weeks after code freeze, which should happen when the milestone here reaches 100%: https://github.com/ethereum/go-ethereum/milestones

AFAIK, ethdev employees have only stated that the milestones need to reach 100% and SEPARATELY that there would be a 2 week announcement when frontier comes out (today's post, I believe)

The idea that the the milestone triggers the 2 weeks was outside speculation (though please post a link if there is evidence to the contrary I missed...)

Great news!

Reply on news.ycombinator.com