The internet is broken
4 stars based on
71 reviews
Thank you very much for that introduction. Alright, well blockchain compression shirts machine right there. Let's try that again. Thank you for that introduction. My talk today is all about this. I'm not going to be talking about this. If you see me reaching for the red hat, please stop me. Before I begin, I want to apologize in advance: I had to blockchain compression shirts what I thought was most interesting. Also I apologize to you, my audience, because this is going to feel like 23 lightning talks back-to-back in quick succession.
I am going to give you a vague taste of things, and you'll have to dig in later. Hopefully I can blockchain compression shirts some of the exciement as I go. I blockchain compression shirts that everyone talking about cryptocurrencies should disclose their holdings. In my case, bitcoin. First we're going to go through a bitcoin keyword primer. This is basically not nenough bitcoin to do anything useful but it does mean you know a few keywords and blockchain compression shirts can sound like you know what you're talking about.
Then wen're going to talk about two kinds of proposed improvements. The first being consensus improvements, changes to the blockchain itself transactions and blocksin particular we're talking about backwards-compatible changes.
And then non-consensus changes, such as changes in how peers on the network talk with each other, which does not involve changing the blockchain format. Today, for a lack of time, I am not going to be talking about anything built on top of bitcoin, layer 2. I am not going to be talking about hard-forks incompatible changes to bitcoin.
So here comes a very quick bitcoin keyword primer. This is a bitcoin transaction. It contains outputs TXOs. Each output has an amount and a script scriptPubKey.
That output script is a very blockchain compression shirts stack-based scripting language with no loops that validates who can spend this output. Here's another transaction spending it. You can see that this transaction has an input, and the input has an input script scriptSig spending the previous output.
In this case, it's a key and the signature of that key. When you evaluate that, and then the output script, it leaves "true" on the stack, meaning it's valid, and therefore that the transaction is valid and authorized. The other term that we come across is "txid" which is the hash of the entire transaction. Transactions are built up into chains like this. Of the 19 outputs on this diagram, 6 of them are unspent. They are considered members of the unspent transaction output blockchain compression shirts UTXO set.
And that's important because to validate a new transaction, the transaction must spend a member of the UTXO set. That's the set of things that you need to check against, to spend a new transaction.
Bitcoin uses a blockchain. A bitcoin block contains the hash of the previous block, causing it to be a chain. And blockchain compression shirts also has a set of transactions in the blockchain compression shirts. Now the blockchain compression shirts that transactions are put into blockchain compression shirts bitcoin block is kind of interesting.
The txid is a hash of the transaction itself. We take pairs of txids, hash those together, and we build up a tree. We put the root of the tree in the bitcoin block header. That kind of hash tree is called a merkle tree. The cool thing about a merkle tree is that I can prove to you that this transaction is in the merkle tree.
If you have the block header then I can prove that the transaction is in the block. But because all I have to give you is the txid and the hash of the other neighbor at that level of the tree, you can combine the blockchain compression shirts yourself, and check that it matches the merkle root.
If it does, then you know that the transaction is in the block. Let's talk about a first set of improvements that have been proposed and researched. These are called soft-forks. A soft-fork is the backwards compatible changes. Old nodes still work. You can make things that are currently legal to be illegal, but you can't make things that are currently illegal to be legal because that would break compatibility. As a warm-up, I am going to talk about an upgrade that was activated in August called segregated witness segwit.
And this addressed the problem of script upgradeability, UTXO bloat, and unwanted malleability. What really is segwit? Blockchain compression shirts, it's a new output script witness type.
And it's literally just a version and then a hash. For version 0, it's either a 20 byte hash hash of a public keyor a 32 byte hash hash of a script. The "input" which normally spends this is empty. The real input data goes into something called the witness. And the witness contains the input stack and the key or script which we promised in the output. You can check whether it's the right one blockchain compression shirts hashing to check whether the hashes match up.
Here's an old style transaction, remember the txid was the hash of the whole thing. If it was a segwit transaction, the input script would be empty and this new witness thing would contain the pubkey and the signature for example.
Now, as you've noticed, that is no longer hashed into the txid, it has been segregated out of the txid. This turns out to be really important because there's a long list of different input scrips that can satisfy those output conditions. You can sign it again, there's a way for third parties to change signatures and make them still work but they completely change the txid, this is called transaction malleability. And malleate is an obsolete word, it means to hit something with a hammer, so I imagine someone out there whacking transactions with a hammer to change their transaction blockchain compression shirts.
Segwit helps several things. We see that version byte helps with-- everything else above version 0, like version 1, is given a free pass at the moment. It helps layer 2, because if you're building things on top of bitcoin, blockchain compression shirts can now rely on the fact that txid values don't change. Someone can't change the chain of transactions by malleating one of the earlier transactions in your pending transaction tree. It also helps hardware wallets because of a checksig change.
But importantly, it helps bloat incentives and throughput. Without segwit, it's actually cheaper to create a new output than it is to spend blockchain compression shirts existing one. This creates incentives to create more UTXOs, which everyone has to remember. With this witness, it's outside blockchain compression shirts old rules-- it's a separate blockchain compression shirts.
These bytes are not counted in the same way. The bitcoin rules say that you cannot have more than a million bytes in a block. The witnesses, which do not apply under that rule, to count a quarter of a byte each.
This means that you can fit more into a block, and in particular what it means is that it's now almost as cheap to spend an output blockchain compression shirts it is to create a new one so the blockchain compression shirts are more aligned.
That's in the past, now. The first of the proposals of the things that I want to talk about for the future is this new problem, in segwit, we have this new output format but we have no address format. The original bitcoin addresses are base58, a nice round number, with a bit sha checksum. It's just blockchain compression shirts example address on screen, don't send money there. The new addresses are base32 with a bit BCH checksum code. Anyway, the code is guaranteed to detect blockchain compression shirts to 4 errors.
And for "similar" letter substitutions it detects up to 5 errors. And if you have other errors, it has a 1 in a billion chance of blockchain compression shirts okaying the result. This is a much stronger checksum than the old bit checksum that we were using.
With bech32, blockchain compression shirts is improved, bech32 is case-insensitive, it fits better on QR codes, and it has error correction and error detection. We're seeing this being rolled out now. I think we're going to see this coming out very soon. This solves the problem of really large scripts. MAST stands for merkleized abstract syntax trees.
When bitcoiners talk about this, what they are really talking about is urning a script into a merkle tree of scripts. Blockchain compression shirts then in order to spend it, you provide one of those scripts, and one of hte hash proofs and the neighbors to prove blockchain compression shirts it's in the tree. Before, we might have had a complex script like this where I can spend it after blocks or you can spend it with your key.