Casper is Pretty Close, Sharding Number One Priority Says Vitalik Buterin

The simple implementation of Casper Proof of Stake (PoS) – which is currently on testnet by two ethereum clients – is pretty close, but not extremely close, Vitalik Buterin, ethereum’s inventor, said at a presentation in Singapore.

There are now two implementations, one in Python and one in Harmony’s Java. Basically things are going very well, he said, and we expect in the next couple of months more Harmony nodes to show up. Hopefully at some point soon, Go Geth and Parity will also implement Casper, he added.

There is still some work left to do. We want to make some small changes, he said, and we are working with academic researchers on formal verification, he added. They’ve suggested some small changes too.

The above does look like a lot, but it sound like they are at the final stage. This might therefore go in with Metropolis Constantinople perhaps by summer.

Yet he doesn’t mention specifics, such as how much you can stake, what would be the reward for Proof of Work miners, for stakers… moving on instead to Plasma.

You’re probably generally familiar with Plasma by now. It’s a second layer that sort of bundles transactions or computations then settles on-chain.

Plasma is happening, he said, without elaborating further on any time estimates. So moving on to sharding after a Plasma overview.

We’re not saying there is one scaling solution to rule them all, he said in the context of sharding. We say… there’s great ideas here, there’s great ideas there, let’s implement them all at the same time.

The difference between sharding and Plasma is that if a block in plasma is invalid, there is an exit mechanism where everyone gets back their money, but the blockchain moves on.

In sharding, it is tightly-coupled. So if a block is invalid, then the whole chain is invalid. And the system makes very sure that doesn’t happen, Buterin said.

We consider any system where rich people run super-nodes unacceptable, Buterin said before adding that sharding’s implementation in a trustless manner is non trivial.

However, there’s a lot of work going on with simple sharding version one, he said, which is already being implemented by multiple teams around the world… the Ethereum Foundation and three others.

You’re probably familiar with much of the general idea. What might be new here is that nodes can not decide what shard they go to, they are randomly allocated to a shard to address different attack vectors.

Scaling is of primary importance, even more important than Proof of Stake, Buterin said. PoS is important, we need to stop wasting electricity like right now, so it’s important, but if no scaling, the only use is token speculation. So I really believe we need scalability so that people can use blockchain applications, he added.

The Ethereum Foundation (EF) is more focused on research, Buterin said. It’s not like Apple where we launch a ready product. We’re not trying to outcompete or monopolize. We believe in decentralization, we want different implementations, different teams.

EF is continuing to expand, continuing to take on new devs. There’s a grants program, so if you want to implement Plasma, sharding, or a smart contracts language to compete with Solidity and Vyper, you can apply for a grant. Will announce stuff about that soon, he said.

His speech concluded with encouraging everyone to get involved in open source protocol work and implementations. If you think you can do it, just do it, he effectively said about a Solidity competitor or a new Plasma implementation etc, and now there’s money for it too in grants.