Are Bitcoin Forks Shitcoins?
https://blossom.primal.net/707d668383755b43feba18e4bd49d77048953c26e9eaa3bf9ed9c79859f015f1.mp4
i'm building a fork and what distinguishes what i'm building over others like luke's effort
- doesn't change the hash function deliberately to make it possible for existing bitcoin miners to switch to it
- changes the difficulty adjustment to a multi-lane scheme that improves the cadence while keeping the same bytes/sats per time ratio, keeping the same emission scheme, while adding a faster block (much smaller, lower reward) that will benefit lightning and ark users a lot. on top of this, each of the lanes responds to high hash power focusing on one or more of the lanes by easing the ones back that are not being hit hard. and as well, all of the adjustment uses a dynamic adjustment system that is well proven in physics (and all of these design elements borrow from physics principles for dynamic adjustment and shock absorbing) called a Proportional Integral Derivative controller (PID controller) - bitcoin's existing scheme is like a clock tick that operates a linear thermostat. it does not cope with big changes in hashpower, and in fact creates a ratchet that makes it such that if the hashpower decline becomes quite extreme (and it is already declining since about a year, for reasons that might be obvious) the chain can dramatically lose cadence, and for which reason the bip-110 fork was doomed to fail, it wasn't going to recover cadence at the 3% of total bitcoin hashpower rate for 460 days, by which time the fork is programmed to deactivate long before (100 days before). it's my assertion that any attempt to fork bitcoin that does not fix the difficulty problem is doomed to failure. bitcoin's security model and difficulty adjustment in fact led to some aspects of its cooption.
- adds a verifiable delay function to the block (in the coinbase transaction) that stops private side-branch mining and reorg/censorship attacks. this is a single thread, serial computation that uses special mathematics and hash chaining that cannot be parallelised. in simulations already built, it has proven to completely stop block cadence from fluctuating very widely at all, it is more important for the short interval low reward low byte size block versions, the long interval, larger block types are simply bigger absorbers of fluctuating hashpower and are harder to reorg anyway - and this is partly why bitcoin stuck to 10 minutes and 1mb.
- it eliminates the witness discount, which means that the more witness in a transaction, the higher the required fee will be and thus as this fork will be parasiting off bitcoin's utxo set, deliberately, it will clear non-spammy transactions a lot faster. this eliminates the primary vulnerability of segwit.
- it bans spending to new taproot transactions, but allows them to be spent into other types of transactions
- it adds a simple hash shielded schnorr x-only bip340 style output called pay to schnorr public key hash (p2spkh) which provides the ability for such as lightning and ark and anyone's private key/signature aggregation to produce a transaction that is in fact from 2 or more parties that looks like a single payment
- it adds a bullet proofs, mimblewimble based confidential transaction - this transaction creates utxos that have no identifiable amounts in the outputs, but it doesn't hide the addresses of endpoints, so it's not quite the same as monero style ringCT, in that it doesn't hide source and destination, but maintains auditability guarantees - it is not allowed as a coinbase output, but for ordinary transaction outputs it is guaranteed it isn't issuing or deleting sats. this is added because confidentiality of amounts in transactions is critical to $5 wrench attack mitigation and is a very good choice for cold storage since you aren't going to spend it often, so it means that you can store a substantial amount of sats in cold storage while nobody can know which utxo it even is that you have got to spend from later (best practices for avoiding correlation via the utxo graph is still something that users have to think about when moving the coins back out of the dark field of no amounts visible).
that's about it. other than that, i really don't want to have it release without at least several dozen node runners running it, because the difficulty adjustment does take a little while to fully hit cadence but it's not long, maybe 5 blocks of each lane so about 40 of these blocks and that works out to about equivalent to about 5 regular blocks. so yes, these early miners will get some substantial reward for working the chain but once it's running it will not be overly generous to early adopters. in fact, because of the VDF, the cadence will not be extremely out of scale as it would be with only the multi-lane PID adjustment scheme, the vdf especially constrains the short/low reward block intervals, and i will be building a proper miner that has existing support for all existing asics so the ideal is that in fact a bunch of old rigs that are currently not economical can be deployed in the early stages and raise the hash power of the network fast right at the beginning.
of course, i will need to write some decent eli5 stuff to explain how the difficulty adjustment algorithm works. analogies relating to dynamic control systems like in segway single wheel electric scooters and as used in jet fighters and multiple other systems (humanoid robots and robotics in general uses them for precision motion control) and the multiple lanes produce a tighter cadence that achieves a more exact emission schedule under any conditions no matter how brutal a large miner attacks the chain with hashpower, and there is no need for checkpoints anymore, which is another lever that such as bitcoin core could use to apply leverage to the chain if hashpower does decline.
it's intended to be a lure for big miners. they will think "no way this thing can survive us blasting it" and then they find that they are earning sats at the same time rate as they do on spamchain. if the market likes it, and i think that a chain that demonstrates high resilience will be appealing, its value will go up and the idea is we lure all the big miners across... and then they have to fight with us about control because all of the levers that enabled the current aggregation of control to central points are deliberately eliminated - taproot, witness discount.
- doesn't change the hash function deliberately to make it possible for existing bitcoin miners to switch to it
- changes the difficulty adjustment to a multi-lane scheme that improves the cadence while keeping the same bytes/sats per time ratio, keeping the same emission scheme, while adding a faster block (much smaller, lower reward) that will benefit lightning and ark users a lot. on top of this, each of the lanes responds to high hash power focusing on one or more of the lanes by easing the ones back that are not being hit hard. and as well, all of the adjustment uses a dynamic adjustment system that is well proven in physics (and all of these design elements borrow from physics principles for dynamic adjustment and shock absorbing) called a Proportional Integral Derivative controller (PID controller) - bitcoin's existing scheme is like a clock tick that operates a linear thermostat. it does not cope with big changes in hashpower, and in fact creates a ratchet that makes it such that if the hashpower decline becomes quite extreme (and it is already declining since about a year, for reasons that might be obvious) the chain can dramatically lose cadence, and for which reason the bip-110 fork was doomed to fail, it wasn't going to recover cadence at the 3% of total bitcoin hashpower rate for 460 days, by which time the fork is programmed to deactivate long before (100 days before). it's my assertion that any attempt to fork bitcoin that does not fix the difficulty problem is doomed to failure. bitcoin's security model and difficulty adjustment in fact led to some aspects of its cooption.
- adds a verifiable delay function to the block (in the coinbase transaction) that stops private side-branch mining and reorg/censorship attacks. this is a single thread, serial computation that uses special mathematics and hash chaining that cannot be parallelised. in simulations already built, it has proven to completely stop block cadence from fluctuating very widely at all, it is more important for the short interval low reward low byte size block versions, the long interval, larger block types are simply bigger absorbers of fluctuating hashpower and are harder to reorg anyway - and this is partly why bitcoin stuck to 10 minutes and 1mb.
- it eliminates the witness discount, which means that the more witness in a transaction, the higher the required fee will be and thus as this fork will be parasiting off bitcoin's utxo set, deliberately, it will clear non-spammy transactions a lot faster. this eliminates the primary vulnerability of segwit.
- it bans spending to new taproot transactions, but allows them to be spent into other types of transactions
- it adds a simple hash shielded schnorr x-only bip340 style output called pay to schnorr public key hash (p2spkh) which provides the ability for such as lightning and ark and anyone's private key/signature aggregation to produce a transaction that is in fact from 2 or more parties that looks like a single payment
- it adds a bullet proofs, mimblewimble based confidential transaction - this transaction creates utxos that have no identifiable amounts in the outputs, but it doesn't hide the addresses of endpoints, so it's not quite the same as monero style ringCT, in that it doesn't hide source and destination, but maintains auditability guarantees - it is not allowed as a coinbase output, but for ordinary transaction outputs it is guaranteed it isn't issuing or deleting sats. this is added because confidentiality of amounts in transactions is critical to $5 wrench attack mitigation and is a very good choice for cold storage since you aren't going to spend it often, so it means that you can store a substantial amount of sats in cold storage while nobody can know which utxo it even is that you have got to spend from later (best practices for avoiding correlation via the utxo graph is still something that users have to think about when moving the coins back out of the dark field of no amounts visible).
that's about it. other than that, i really don't want to have it release without at least several dozen node runners running it, because the difficulty adjustment does take a little while to fully hit cadence but it's not long, maybe 5 blocks of each lane so about 40 of these blocks and that works out to about equivalent to about 5 regular blocks. so yes, these early miners will get some substantial reward for working the chain but once it's running it will not be overly generous to early adopters. in fact, because of the VDF, the cadence will not be extremely out of scale as it would be with only the multi-lane PID adjustment scheme, the vdf especially constrains the short/low reward block intervals, and i will be building a proper miner that has existing support for all existing asics so the ideal is that in fact a bunch of old rigs that are currently not economical can be deployed in the early stages and raise the hash power of the network fast right at the beginning.
of course, i will need to write some decent eli5 stuff to explain how the difficulty adjustment algorithm works. analogies relating to dynamic control systems like in segway single wheel electric scooters and as used in jet fighters and multiple other systems (humanoid robots and robotics in general uses them for precision motion control) and the multiple lanes produce a tighter cadence that achieves a more exact emission schedule under any conditions no matter how brutal a large miner attacks the chain with hashpower, and there is no need for checkpoints anymore, which is another lever that such as bitcoin core could use to apply leverage to the chain if hashpower does decline.
it's intended to be a lure for big miners. they will think "no way this thing can survive us blasting it" and then they find that they are earning sats at the same time rate as they do on spamchain. if the market likes it, and i think that a chain that demonstrates high resilience will be appealing, its value will go up and the idea is we lure all the big miners across... and then they have to fight with us about control because all of the levers that enabled the current aggregation of control to central points are deliberately eliminated - taproot, witness discount.
4❤️1
