The bitcoin fork wars continue with the newest one, Bitcoin Gold (BTG), testing Schrödinger’s cat by launching and not launching at the same time.
The countdown is on, with the currency to fork at block 491,407, that’s around October 24th, just two days from now. But while it will fork at that block number, it actually won’t fork on that day.
Lead developer of Bitcoin Gold, who goes by the nickname h4x3rotab and appropriately has a cat avatar, told Trustnodes earlier today:
“When bitcoin reaches the fork height, bitcoin will continue as if nothing happens. At the same time, our client is not ready, so there’s no main chain of BTG. But after a few days, when BTG client is ready for mainchain launch, the BTG blockchain will be created.
The BTG blockchain starts from the fork height and then the premine begins, coming with public mining. So as a result, BTG starts from the “snapshot” crated when bitcoin reaches the fork height.”
In effect, this is a backwards chain-split. Instead of creating a new currency at the same time as the block is reached, the project will only “freeze time” there.
After a few days, or more likely weeks, when they are actually ready to do a real actual chain-split, they will discard all that history accumulated since October 24th, with their clients so synchronizing up to block 491,407, and then diverging to their own chain.
The given reason for this roundabout way of forking while not forking being that they decided on the block number and are unable to change it now.
But the project does not appear to be anywhere ready for launch. Our time estimate is months, or at least weeks, although they tell us days. Perhaps.
Their code is open source, and Bittrex, a crypto-exchange which lists a vast amount of digital currencies, had this to say about it:
“Bitcoin Gold does not satisfy our criteria for safety for our users,” after stating that Bitcoin Gold does not have “fully formed consensus code, implemented replay protection, adequate code for testing and auditing, publicly known code developers.”
Martin Kuvandzhiev, a core developer for the project, confirmed the code aspects, telling Trustnodes “we are currently not ready with the code.” He further publicly said: “We are working on the code still. The [testnet] is syncing. But very very slow.”
H4x3rotab says “when essential features like anti-replay is done, we will be able to bring up the testnet.” He also confirms the consensus code has not been finalized, stating “currently mostly focusing on replay protection. And the bootstrap of the testnet.”
So quite a bit more work to do until the actual, real, launch. However, even that won’t be a real, real, launch because they are to have a pre-mine window before opening it to the public for mining.
They have stated they will pre-mine no more than 1%, or around 160,000 BTG. Kuvandzhiev says “75% locked for the next 3 years so the development can continue. Most of the money are for bounties for developing new features. The core team will have the money on multi signature wallet and the money will be released each month for development.”
H4x3rotab further clarifies that 75% of the premined coins will be locked for “36 months. So these coins will be separated to 36 pieces equally and get unlock monthly.”
Projects usually don’t release the full code when there is a pre-mine, but in this case h4x3rotab is adamant that the open source code is actually the entire code. Explaining how a pre-mine and full open source code prior to launch can co-exist, he says:
“We define a range [fork_height, fork_height + premine_window). The consensus rule enforce that during the premine window, all blocks have 1 diff and the output address of the coinbase tx must be whitelisted.
The whitelisted address are mult-sig wallets controlled by the founders. Another consensus rule enforces 75% of the premined coins will be time-locked for around 3 years and get unlocked every month.”
This project is founded by Jack Liao, a Chinese miner, with the aim of creating a more decentralized bitcoin version through an intentional minority chain-split by changing Proof of Work to an ASICS resistant algorithm.
While the name clearly hints at the settlement network with Kuvandzhiev stating they are “friends with bitcoin and we will be glad to work with them side by side,” in response to whether this was Bitcoin Core’s back-up plan.
The pre-mine might raise some doubts on that since quite a few Bitcoin Core developers have criticized other projects solely on the basis of having a pre-mine, but would simply changing Proof of Work make bitcoin more decentralized?
There are two main aspects of centralization in PoW mining. One of them is pools, which as the name says pools many miners into, effectively, just one node. The PoW algorithm is irrelevant to that aspect.
Another one is ASICS, customized special purpose hardware that’s specifically designed for mining. They are far more expensive than GPUs, and since they have only one purpose, less people have them.
But it isn’t necessarily the nature of the hardware that creates centralizing pressures. It is rather the fact that the more you have, the more you have a share of the network’s newly issued currency.
As such, the nature of the hardware is in many ways irrelevant because instead of rows and rows of ASICS farms, there would be GPU or even laptop farms. And while there aren’t many producers of ASICS, GPUs also have only two main producers, AMD and Nvidia.
So any decentralization would most likely be temporary until big GPU mining farms – which already are in place to some extent – take most, if not some 90% or more, of the network’s share.
Moreover, and very much counterintuitively, GPUs might actually make the currency more centralized because the network wouldn’t easily be able to escape the GPU miners. While with ASICS miners, you can just change Proof of Work.