Topic: A blockchainless Digital Collectible Currency competes with digital renminbi (Read 136 times)

Hmm rethinking the difficulty settings.  First of all I don't think pinning the value on something else is a good idea because I don't think we can count on knowing the price of our coin by looking at exchanges and adjust based on that since those might be regulated away at some point, and even if they are not we should try to become a stable commodity in ourselves because that adds value to what we are doing here.

So my thought now first and foremost is that nodes will be the ones "bidding" on the bit-length of the factors by what they accept.  And they should be allowed to bid whatever they want.  For example they could say that this month they accept any prime factorization of 500+ bit numbers (this is in binary not base 10.  Binary is best as allows most adjustability of difficulty).  There will be a bidding war of the nodes and so the lower they set their bit requirement, the more customers they will get.  However this also means they are diluting the value of their staked coins so a balance needs to be struck.

My suggestion is, and the customers will ultimatly be the ones that have to enforce their morals on the nodes by choosing who to do business with, is as follows.  The minimum bit length accepted by a node should be one that allows a mid-level consumer grade enthusiast PC (so here in 2021 a good i7 or Ryzen 7) with a decent graphics card roughly the same price as the processor, would be able to find a solution once a week working at full capacity by themselves.  This means that low end and mobile devices could maybe hope to win one award a month, which is ok for a casual miner.

So anyway this is just an idea and really the nodes, miners, and coin users would be the ones reaching a pseudo-consensus on how the difficulty should be adjusted.  But if people kind of set an expectation that a good PC should win 1 award a week on average, I think the price should stay pretty stable and the coin will keep a wide appeal because people feel like they are getting fairly rewarded for their work.  This also should work since GNFS has shown to be very resistant to speedup so should only be effected by moores law just like consumer pc's.

Also my suggestion is that nodes rethink their difficulty requirements every month.  In the US, the constitution was supposed to be a living document but the requirements to change it are so high that very few times has it been changed.  I think constitutional conventions should happen yearly and this would have avoided the country in its dire state right now.  So since this coin is a living digital constitution, I suggest likewise that difficulty adjustments happen every month so that a runaway problem like inflation or deflation doesn't happen.

I also suggest that the difficulty should not change more than either 3 bit length either up or down per month.  So if past month was 500+ bit length, next month the requirement should be 497+ or 503+ or somewhere in between.  Obviously this is just a suggestion and if there is good reason to go a little outside these bounds then that is a decision each node makes, and ultimatly each customer of the nodes.  Miners, nodes, and customers all need to make their voices heard and vote with their work and money and their feet so that everything stays fair.  And I think commiting to reevaluate difficulty monthly or at least yearly is essential to keeping things fair.

I'm thinking that instead of pinning the value on some relatively stable commodity (just realized that there is a maple syrup mafia that sets the price for 70% of the worlds supply lol) we can be a stable commodity ourselves.  I think if we extrapolate with moors law in the speed of factoring large numbers, we can set a roadmap where every month or so the length of number to factor grows by 1 digit.  I will work on calculations to determine some sort of difficulty schedule.  

Otherwise if we do want to pin to a commodity, then silicon would be a good one since it is used in solar panels, computer chips, and even electricity generation - which all fits well with a cryptocurrency - and it can be made from dirt anywhere in the world with just heat and charcoal. 

Honestly I think pinning to silicon is best come to think of it.  We can guess to try to make the price stable using moors law but sooner or later each further digit would just get exponentially harder and a supply crunch will ensue.  Currently the design is pin the value to silicon spot price per lb (which is currently around a dollar, so works well since this is the smallest trade-able unit). 

Setting a low price also may or may not guard against quantum computers.  They will have to know the state of every qubit before running shors algorithm, so if they only can get a dollar for their trouble, might not be worth it.  But who knows.

Just alerting people that this thread has been moved by the mods from bitcoin technical discussion to altcoin discussion.

Great questions and I think you are honing in on the weakest part of this which is consensus.  However I do want to preface that I just read wei dai's b-money and in his idea he also concluded that a "makeshift" consensus like I am proposing could work.

So in terms of inflation we can choose whatever we want and my thought would be there is a value target.  We can pin it to the dollar or pin it to gold or maple syrup (not even kidding on that last one).  What would happen is we would see the market price, and we would increase or decrease the bit length requirement of numbers to factor to be accepted on the network.  Say adjustment is every month.  If last month a coin was trading for 1.2 pints of maple syrup, then instead of say a 200+ digit number to factor, you will have to submit the solution to a 199+ digit number (to try to get value down to 1 pint of maple syrup) to be accepted by the network.  Also we could say that there can never be more than x change to bit lengths in one month.  So say that number is 2, the most the difficulty could be adjusted in a month is 2 so 198 or 202 are the limits to what it could be next month.  This way there isn't much overshooting and price stays pretty stable.  

So it would be an update to the node software, and of course nodes could either update or not.  The network can fork and if that happens fungibility of future mined coins might be reduced, harming everyones value so the incentive would be for nodes to reach as broad of consensus as they can, while still maintaining their morals.

It is always good to reconsider technologies from roots up, both for innovation and for educational purposes, so, I liked the spirit and the style, congratulations, op.

I need to  read your articles for further discussion but as of now I got to questions to ask:
1) What is the inflation policy?
2) Without a consensus mechanism, how such a policy is enforced?

Thanks for looking through this.  Ledger does not equal blockchain.  A ledger in this case is just a database like an excel spreadsheet that says who owns what.  A blockchain is not required for there to be a ledger.

Yes you are connecting to an ip, but the ip is constant and can develop a reputation.  Also that ip could have coins staked that would be destroyed on every other nodes ledgers, if they misbehaved.

They get rewarded by you paying them to facilitate key changes (transactions).  I see it like podcasts, you pay one service to host it and it gets picked up on other distributers by rss feeds.  So lets say a popular node says that for 1 coin a month you can send them all the key requests you want, then the other nodes pick up these transactions from that node.

Running a node is like running a website, not like mining.  It does not require hashpower or large power usage, it just has to serve data like a website does so it could be run on a raspberry pi for super cheap.

Miners get compensated by creating their own proof of work problem, and solving it.  If the problem meets the requirements of the nodes, they will list the solution the miner found as a "coin" that only the miner owns the private key to.

But you still have to trust someone, so why not trusting the majority of hashpower instead? What do you mean by saying "you choose who to trust"? Aren't I connecting to IP addresses that are unknown to me?

I'm confused. Is there gonna be a ledger or not? Because, at the beginning of your post you are suggesting a solution without a blockchain. If there's going to be a database, then blockchain exists.

After all, I still don't understand what are the incentives of running a node. Just a fee that is chosen by the user? I also don't understand how miners get rewarded on DCC. I think that there are parts of this solution missing.

I want to get all of your takes on the validity of an idea I have been hashing out for several years now and finally have it in a form that I think would work.

Satoshi designed the blockchain to solve many technical challenges at once including:

• creating coins
• distributing coins fairly
• creating a ledger
• transferring coins
• preventing double spend
• ensure everyone can compete for coins with consumer hardware
• converging the ledger to one version

What if we could create solutions to all or nearly all of these with no blockchain?  Without a blockchain and thus no time-gating, then transactions can be near instant, mining can be done without connection to the internet, 51% attack impossible, and price can be stabilized.

The solutions I have come up with is called a Digital Collectible and a currency based on it called a digital collectible currency (DCC).  This digital collectible is an indivisible solution to a problem created from a private and public key.  Basically a miner would create a private key and public key for each coin.  Then they would hash it with a nonce to get a large random number. Next they would factor this number using GNFS.  The list of prime factors is the Proof of Work and is the Digital Collectible.

At this stage the first two problems, creating and distributing coins, is accomplished.

To create a ledger we would have many nodes each maintaining a database and recording everyone's public key, nonce, and factored solution.  Each line in the database is a "coin".  They can use GIT or something to show changes over time to their ledger.

To transfer coins the private key of each coin to be transferred would be given to someone else.  There is no direct transfer from one public key to another.  Now both the giver and receiver of the coin have the private key, and thus either one can spend it, which is the double spend problem.

To solve the double spend problem an update request to the ledger can be made.  Anyone with the private key can change the private and public key of the coin by sending a message to every node requesting a public key change that is signed with the original private key and which integrates a new private key for the new public key.  So to validate that this transfer happened, you ping the network of nodes and just confirm that indeed their ledger now has changed the public key to the value you proposed.  Which message was received first, the one where you change the private key, or the person that sent it to you changes the private key on you?  Each node decides that.  And by the receiver pinging the nodes and finding out which transactions their trusted nodes thought was first, the receiver can verify that indeed they were first and they now own the coin.  If the other person was first, then the coin receiver can decline to give the product to the coin giver, and have him send a new private key.

This idea using GNFS factoring large numbers is the best possible way to ensure consumer hardware can mine it and it does not benefit from GPU, FPGA, or ASIC speedup.  However since it is an NP-hard problem and not NP-complete like the hashing algorithms, it can benefit from quantum computer speedup.  However a quantum computer that can factor a number as large as a desktop can today, would take decades of development.  Also a quantum computer would not win every coin nor could it attack the network, since other coins could be mined by other people simultaneously.

The only problem yet to solve is converging the ledger to one official version.  But I think this is actually better, not having an official version would make the ledger flexible and changeable by consensus, and immune to a 51% attack.  We would have to trust nodes, and there can be things done to increase trust such as having nodes stake coins, and blacklisting any misbehaving nodes.  Nodes would be in a constant war of winning your trust, and you would not have to rely on any one node and can send your requests to as many as you like and nodes would be incentivized to cross check with each other to ensure their ledger is trustworthy.  Nodes could require you pay a fee to send them your key change requests (transactions).  I foresee it kind of like sharing a podcast.  You pay one service to host it, and all the other services pick it up from the RSS and host it for free.

Even though bitcoin claims to be trust-less, in fact you are forced to trust the majority of hash-power.  In this idea we do not claim it is trust-less, and yet you are not forced to trust anyone or group, you can choose who to trust.  The customer having this choice of nodes, as a free market, is what makes the network robust, and nodes would have to win customers by being more trusted than their competitors by developing a good reputation and connections and/or staking a lot of coins that would be lost if they misbehave.  A node is a public IP address, and when a node pings them, they need to sign for their staked coins.  If they refuse to sign, then it is assumed they have no staked coins.  If word gets out that this IP is trouble, those coins would be blacklisted from all other good nodes.  Even if a large percentage are bad, the network can restructure eliminating the bad nodes and no value or security is lost from the coin since nodes are not contributing hash-power.

There can also be a list that nodes hold of Node Staked Coin public addresses (NSC addresses), and also Blacklisted Addresses, so when someone is offering you coins from a public key for trade that is on this list, to be wary that they may be trying to offload their coins because they see judgement coming from the other nodes.  These coins would be lost forever if they are blacklisted, no matter who the current owner is when they get blacklisted.

Here are archive links to some articles I wrote on this with more information:
https://archive.is/T3P0I
https://archive.is/zDOEP

Again I would ask you to try to find holes in the logic and let me know if you have ideas to improve it.  Thanks.