 OK. And after it is verified by nodes, if there is no accident, it will be included in the blockchain and the block. So the block has the field, previous block hash, and timestamp, and the transactions. And the blockchain is actually the chain of blocks. And it is chained by the hashes. So let's see this blockchain. The block A, block B, the block B's block hash is set to the block A's hash. It is chained. And it is append only. So if you want to make a new block, you must append to the existing block, blockchain. So assume there is a block C. You must set the previous block hash as the latest block. And it will append it to the existing blockchain. And the purpose of the chain of blocks is it makes attackers hard to tamper the blockchain. Because if you modify any fields, anything in the blockchain history, you can find out that when you do calculations, when verifying the whole blockchain. And a transaction is confirmed only if it is recorded in the blockchain. And the order of transaction is determined because inside one block, there is an order of transactions because it is a list. And between the blocks, because you know the order of the transactions, sorry, you know the order of the blocks, so you can determine all of the transactions order. And this blockchain is used to serve as the single history for all nodes. So all of the nodes just believe in this single history. And there is some question like how to extend the blockchain. Actually, everyone can extend the blockchain. But the one with more power has the bigger chance to extend the blockchain. So the incentive of doing this is money. It is so-called mining. Let's assume there are two people here, two nodes. And one is bigger, one is smaller. So the chance to extend the blockchain is bigger for the one with more power. And let's define power. In Bitcoin and the current Ethereum, we use proof of work, which is known as POW. And this is the one with more computing power, has more power. And another famous one is POS. This is the idea is the one with more stake has more power. The stake means how much money do you have in the blockchain? Yes. And let's talk about Ethereum. Ethereum, it has its own blockchain. And you use POW. And we will change to the POS in the future. Maybe some people already heard of the keyword Casper. It is still in the testnet testing. And another feature of Ethereum is it has a smart contract. So the right-hand side is co-snip it, a snippet in Viper. So it is the basic contract that maintains the counter. And it has a method that allows the user to call. You can increment the counter. And then you have a method to get the counter's value. And currently we have a scaling issue. Many people are concerned now because the current transaction per second is slow. Like Bitcoin has a TPS of 7, and Ethereum is 15. And while the Visa has thousands of TPS. And there are more and more applications appearing in Ethereum, like the tokens known by everyone, ERC20, and CryptoKitties. It is the famous game. And there is one period that the transaction of the CryptoKitties stuck the whole Ethereum network. So make the payments hard in Ethereum at that time. And CryptoDonb and other games. So we need a higher TPS. And what is the reason causing this? Actually, one of the reasons is every node handles all of the transactions. So if we want to deal with the problem, we must make one node only deal with part of the transactions instead of all of them. So it is the idea of sharding. About Ethereum sharding, the basic idea is we just divide a big blockchain into multiple shards. So think of it as a really big, big blockchain. And transaction processing capability of the main chain. And we have the shard count is big OC. And transaction processing capability of the shard chain is also big OC because it is a chain. So the total of the whole sharding system will be shard count times. And I have a scenario of the collater how it works. And the scheme I just mentioned is the old spec instead of a new one. The new one is becoming more complicated. So I just introduced the old one. And the phase two is, oh, sorry, the phase one, basically we don't care about the execution of transactions. And until phase two, we edit back, which means we allow transaction executed. And phase three is like high on the state protocol. And phase four, we will implement the cross shard transactions. This is actually a hard problem. And we're still figuring out how to do it. And the phase five, the type coupling with main chain security. Because as I just described, we're doing something like we depends on the main chain. That is loose coupling. So everything in the shard chain is determined by the main chain. So in this phase, we will change that. And the phase six, we will achieve super quadratic sharding. And that's my talk. So the reference from Bitcoin, White Paper, Ethereum, White Paper. And the sharding phase one spec is in our forum. It's a post. So if you're interested, you can check it out. And then the sharding FAQ, it describes the problems we encountered. And we're still discussing about. And the sharding repo, it contains the old code in our sharding POC. And we will put our new contract there and the documents there. And the current implementation in pie EVM is this repository. That's it. Go back to the roadmap. OK. So Kabi, you are working on which phase now? Phase one. Phase one. Yes. Then the last phase, phase six, your target completed year, which year you complete? Target to complete a whole phase of six. Sorry. Can you repeat that? My question is the whole phase is six phases, right? Yes. So your target always or target complete all the phases? Or is it timeline? Timeline? Yeah. I think it is not clear for now. Yes. Because still progressing, something are still under discussion. Yes. So how many people are working on this with you? You mean sharding? Yeah. Like five people? Yes, currently. And some of us are in Taiwan. And some of one researcher is in London. And Vitalik will give us a lot of suggestion. Is it like an open development model or are you? It is open source. What's this part of Q? Can you go back one slide? OK. So I know the one that had the URL on it. URL, OK. So look at the third URL, sharding phase one spec. That research board is where you can sort of see it little by little. And if you want to get more working on it and talk to them, you get. Yeah, if you stand, you'll be. But what's your first thought? There are actually a lot of posts discussed in the whole spec and the problems. Sorry. So inevitably, there's going to be like shards that will be executing transactions that will be consistent, right? You mean cross shard? Or shard A and shard B might have transactions when they resolve to maintain my conflict? My conflict? I don't think so. It should be separate states. So the execution within one shard does not affect it by another shard. What if the execution in shard A empties all of Alice's account and execution in shard B tries to take money out of Alice's account as well? Both of those can't get part. For my understanding, we separate the state previously. So if we want to do some cross shard things, we will go back to Manchin. And the consensus will be made in Manchin. Is it that the code that splits the shards in the first place guarantees there's no conflicts conflicting accounts ahead of time? Or is that resolved when you try to reconcile in the main block? You can think that in our early stage, if you want to do some operation in some shard, you have to lock your money in Manchin. And therefore, you can create some money in that shard. So the operation will not be affected. How far are you into the phase one implementation? How far? Well, where would he manage to move in? It is a good question. Because we just have a new spec. Yes, so many things still need modifying a lot. Yeah. Could you describe the other things that are being done to improve the transaction state? Oh, OK. So for sharding, it is a layer one solution. It is down inside the blockchain. And there are still other options, like layer two solutions. It is on top of the blockchain. And there are state channel. And there are plasma. And like state channel basically make many of the transactions offline. So many micropayments transactions, they are not going to the blockchain. So it dramatically reduced the number of the transactions. Yes. I know some of that. And so there is a nice post on medium. It's called making sense of a serious layer two scaling solution. State channel plasma throughput. So if you just google plasma space throughput, it will probably get it. And it's pretty good. Thank you for sure. So I can see how this would increase the volume of transactions that we have been processing. Yes. But the introduction of breaking out the shards and break it down into shards, is that in practice is that going to increase the latency? Like going to make it take 10 seconds. It's not going to be 15 seconds. But instead of 1,000, we can do 30,000 transactions. Is it going to increase the latency of the transactions? latency of transactions? You mean within one shard? Or? Well, ultimately what counts is that it gets to the final blockchain. So the time from introducing a transaction to it being on the final blockchain that everybody in the world can see it, right now it's approximately 10 seconds. With sharding, I see that it's going to give you a larger volume of transactions in that period of time. Will it also increase the latency that it's now taking 15 to 20 to 30 seconds before it shows up? Or are you still doing that for the 10 second time? OK, so in the sharding phase 1, one period has one collision. And the current period is five blocks in main chain. So there are five times. The current block time is about 15. So it's 75. So in these settings, the latency will increase. But there are still some research about this, like reducing the number of blocks within a period, maybe reducing it to one. And it will make the latency just like the one in original Ethereum. Yes? So in general, most people increase latency. There are a couple tricks you can do to hydrate back and remain a little ambiguous how people are going to help surface switching to the state. We'll see if you can see that. But I mean, in short, there's a lot of trade-offs for these things, whether it be like block latency, how many conductive do or some finality in decentralization. And in general, our road map is we're not trying to match that on any one of these. We're trying to be kind of decent and normal. Obviously, if you have a system that you want to be real, say you want to play Starcraft and boxing, you will probably never be like for you. Yeah, and we've just worked with this as a design system that we're trying to do decent on with all of these. Yeah, I don't think the best thing to do, because I think it certainly has a real record, right? I think we want the volume to be real hard enough that we're getting it soon. Yes. Any more questions? OK. Any shots, do you think, about 800? Currently, in our phase one spec, it is 100 currently. But it should be larger. It can be extended. It's because 100, the latency increased by 100. Anyone give us shots? You mean latency? Or? OK, just like you said, the latency increased. It is not because of, it is not due to the short count, number of shots. It is not because of that. It is because we set the period as five blocks in mention. Yes? Ethereum, is this just going to magically happen that you're using a sharded system versus the main block? Or like for phase one, it sounds like I'm referring that wallets will choose whether to go into the sharded system or whether to go and wait on the main system. So it's an active choice now? Any idea, Sam? How do you know if your transaction is going on a shard or not shard or if they're a choice or if it's happening? I'm not so sure of it, actually. Yes, but. You have shard in working now. I think we'll get some transactions that go through shard. You're going to end up on the blockchain now, right? Yes. So how do those transactions get on the shard if they had to choose to do a shard in contract? If a node wants to send transaction to a specific shard, it needs to listen to the network of that shard. So the node makes the decision? Yes. The node makes the decision. So you can actually choose which shard which you are interested in. When you say phase one doesn't implement the EPM, does that mean it only implements the transactions but no smart contracts? Yes. In phase one, we just treat the transaction data as the data. So we don't actually execute them. Because we want to focus on the data availability problem. Yes. Any more questions? OK, thank you.