The real problem is if we did as you suggest then users would run their node only for transactions
and as soon as it is broadcasted and within a block, they would just shut down that node.
The real discussion by the OP is about having 24/7 full nodes that help distribute the security and
risk through global decentralization which is encouraged by directly incentivizing the transaction node
users. Miners are doing their jobs 24/7 and not intermittently. Nodes should be doing their jobs 24/7
as well, like guards on distant mountains keeping watch of their territory.
Average joe users do not need to nor should run full nodes, it is beyond their capabilities, but
full nodes should be incentivized in some way to encourage new capable people into the full
node field, so we can get the Bitcoin node network from 5,000 to around 50,000+.
An increase in legitimate 24/7 full nodes distributed worldwide is the goal.
Nodes are incentivised to stay online since if they restart, they will have to regenerate their NodeID and re-index the block chain. After an initial index this will be until at least 1 old checkpoint and at least 2 new checkpoints have been verified. The return will not be instantaneous. |
|
|
|
Yes it is-- the circuit it implements is universal, meaning that it can emulate any other circuit.
Then by your own definition it is not "Application Specific". I think we will just have to agree to disagree on this one. |
|
|
|
Yep. (and if it did-- what would you run it on, your CPU is an asic too!)
A CPU is not an Application Specific Integrated Circuit. It is application agnostic. |
|
|
|
This would just increase the digital divide, people with a poor internet connection and little hard drive space to spare would not be able to participate.
One of the nice things about bitcoin is its accessibility, an impoverished farmer can walk into a dilapidated shed, connect to the dial up connection with the crash-prone 2003 computer, and spend his hard-earned milli-bitcoins on a sack of grain.
Requiring users to run a node would turn bitcoin into an elitist system that can only be used by people with fancy computers, fast connections and large hard drives.
My PoN idea gives a privilege to node runners, but still lets others use the system, they just have to wait for their turn while the node-runners skip to the front of the line.
I thought the argument against this was "Moores Law means technology will be dirt cheap"? At least that was what everyone keeps saying. However. I have hinted at a distributed block chain that would mean an on-disk size less than 1GB regardless of chain size meaning the concept of a full node as one that has to have the entire chain is irrelevant. |
|
|
|
Having an access point is a requirement to participate.
The problem is exactly the negation of that statement, you don't need a node to use bitcoin, there is little incentive to run one (I am not saying that everyone should be required to run a node to use bitcoin, that would reduce the accesability of using bitcoin). You misunderstand. I am proposing that we make it a requirement to participate, just like the access point. Then all wallets are supportive of the network and the argument about finance vs altruism is moot. |
|
|
|
-Either both miners and node dont get paid = voluntary work in the name of altruism
-Or both get paid = for profit business
You cant mix them, as it is contradictory.
This is a false dichotomy. When one wants broadband to the home, one requires an access point. This is in addition to the LLC, fibre and all the infrastructure owned by the provider. Having an access point is a requirement to participate.. They don't pay one for having an access point. The same argument can be made for the required parts of the Bitcoin system and, indeed, it was with the first implementations where one had to have mining, full node and wallet. So the choice isn't altruism or financial reward, it is a requirement to participate or financial reward for infrastructure support. |
|
|
|
True.
However, I see "ASIC resistance" used more as a blanket term for the phrase "Is it more cost effective to mine my coin using general hardware as opposed to making an 'Application Specific Integrated Circuit', whatever that 'circuit' may be" and its variants. That is, a purely economic term. Of course, if you subscribe to the supposition that a "specific circuit" will always be better than a "general circuit" for the specific case it is designed for.
Cost has nothing to do with it. The meaning is established. ASICS scale by parallelism. If an algorithm is atomic and memory light then ASICS are a good choice. They only have a fixed number of gates and creating memory out of those gates uses a lot of them. As an example. The current hash calculation needs only the hash states and a simple counter for the on-chip memory which can then be applied to a non-changing memory of transactions-a constant, if you will. This can easily be scaled to multiple hash calculations in parallel on a single memory list of transactions. If, instead of a counter, the hash was calculated on a varying number of transactions (1 then 2 then 3 and so on) then this is not the case. Scaling up the hash alone yields the same hash if the transaction list is constant and in order to calculate the hash for each (1,2,3...) transaction lists would require a huge amount of resources which the ASIC is unlikely to have. This method would be ASIC resistant. |
|
|
|
It was a basic assumption of satoshi that mining will remain decentralized since he assumed that 1 miner will be 1 node, and everyone will mine on his PC with the CPU.
This is incorrect. Satoshi had this to say on the subject. The current system where every user is a network node is not the intended configuration for large scale. That would be like every Usenet user runs their own NNTP server. The design supports letting users just be users. The more burden it is to run a node, the fewer nodes there will be. Those few nodes will be big server farms. The rest will be client nodes that only do transactions and don't generate. (Sauce)The scenarios are playing out as expected with an end game of a cartel of miners, users with web wallets and the occasional hipster running their own full node. The idea of economic incentives for full nodes is an attempt for already committed early adopters, to squeeze a few crumbs from the miners' table in the face of the inevitable. In the end, most full nodes will be run by the miners so it is a zero-sum game. |
|
|
|
A "true rollback" (which as I said isn't being used by an major RBDMS that I am aware of) would actually be much faster than any other method (i.e. you either are going to have to issue update queries to act as though things are being unwound or restore an earlier backup which you'd then have to perform a partial log restore from).
Understand also that even if deletes involve appending the data to the log you would still be able to truncate the log for "checkpoints" (at which point any of that extra wasted space is recovered).
SQLite enables one to create " Savepoints" which are like GOTO statements for rollback. |
|
|
|
What's guaranteed to happen is that if you double the blocksize right now, we'll see nodes dropping faster than allies in normandy.
Unfortunately, there is no way to scale anywhere notable on-chain, it's just not realistic, unless of course, you don't mind bitcoin being hosted in centralized hubs, which would kill the point of bitcoin.
Objectively, segwit is the best thing we have now. Those not supporting it are simply not aware of the realities we are dealing it, or they are, but don't care that bitcoin's network becomes centralized.
Sooner or later, segwit will be actives. Even if the morons push a blocksize increase, they will quickly see the blocks full again and realize the futility of scaling on-chain through simple blocksize increases without other methods.
It's already hosted in centralised hubs. That boat sailed a long time ago. Now that no one cares about the on-disk size, why not make it 4 times bigger so that online wallets don't have to fix their database queries? Block size is a distraction. Now the economists have made a scarcity market of the spam prevention, they will let that be resolved when hell freezes over. |
|
|
|
|
When an industry is giving out billions in bonuses while everyone else has to visit foodbanks; cost isn't a significant factor especially when the public purse will pick up the bill if it all goes wrong.
The idea that the banking industry will be disrupted by this technology is almost laughable. Especially since they have the capability to suck the few knowledgeable engineers, if they haven't already, out of the resource market and negotiate with the authorities to accept their alt-coin as tax payments. All that remains then is let the miners fight it out for dominance and take over the winners. Bitcoin (or it's alt) then becomes the inter-bank transfer service for VISA and Mastercard since the users are treated just as cash cows to the tech without a voice on the chain; they won't even notice.
The Chicago School have their fingerprints all over this, now, which is why decentralization is resisted so vehemently and the economists micro manage the software direction. It is starting to look more and more like Bitcoin is the dieing gasp of a rejected economic school as the world lurches to the right in a misguided attempt to get rid of them.
It did not have to be this way.
|
|
|
|
If you're having problems with understanding that, Bitcoin isn't for you right now. You'll just have to wait until people that are capable of understanding these simplicities have established your faith in the change.
Unhelpful elitism. |
|
|
|
How much easier or harder does SegWit make transaction data mining and address tracking?
Not at all. Segwit does not make it any harder or any easier to track spends. But isn't transaction malleability a barrier to tracking? So much so that exchanges can't keep account and therefore want SegWit to remove that barrier? |
|
|
|
|
How much easier or harder does SegWit make transaction data mining and address tracking?
|
|
|
|
I appreciated the responses to my question TransaDox, I've learned a fair bit. The conversation is out of my pay grade, so I've had to do a lot of side-reading to make sense of it.
I'm not sure why you got attacked for it. Who knows if maidsafe will pull it off? It certainly feels like an esoteric challenge to understand it all. And far too complicated to be dismissed out of hand. It's several new layers of internet protocol after all, it doesn't play by the same rules and the deeper I look the more fascinated I become.
Domain experts tend to be very focused on their sphere of expertise to the exclusion of all else. Therefore the forum style of conversations where there are sporadic wanderings - akin to sidechains in bitcoin parlance - tend to get conflated and perceived as noise to their message. Some are more tolerant than others of these interruptions and partake in the side conversations to entertain and explain to non-domain experts. At the other extreme; others lambaste as "you don't know what you are talking about" and deem it beneath them to explain, entertain or teach. This thread is somewhere in between. I wouldn't get too bogged down in the Maidsafe implementation. It is a monetisation of the blockchain based on the work of IPFS. If one reads the well written IPFS document I linked to, it gives a better overview of the technology. Maidsafe has used an alt-coin as a method of account for the BitSwap Protocol (Section 3.4). I also suspect Kim Dotcoms new platform will be something similar to BitSwap (if not exactly). |
|
|
|
That is a very nice graph. It does a good job explaining why this shouldn't be so much of concern. SPV clients make it easier for the average joe to spend btc and nodes are the only ones that will need beefy drives. Only Moores law isn't linear so that graph is misleading. Don't be fooled by the deniers. Moores law is an observation, not a "law", and the tech industry is already downplaying its significance. CPU speed has plateaued and memory (which is the tech being flouted by the 700GB fanantics) hasn't been following Moores law for quite some time. Physics is a bitch and if one looks at all the technologies then the industry has come to the conclusion that Moores law is no longer valid. Intel, recently, changed Moores law to a "design goal" rather than a rigid expectation. The shamen praying to Moores law for future proofing the blockchain will be finding their incantations insufficient. |
|
|
|
But it does contribute to the network. It still has outgoing connections and it still validates and then relays all blocks and transactions that it receives to its peers.
The distinction should be that a full wallet is a full node, but not the other way around. A full node fully verifies and validates every block and transaction it receives and then relays them to its peers. However a full node need not necessarily have wallet capabilities. A full wallet is a full node with wallet capabilities.
But it doesn't need 100GB of disk storage to do that after it has synch'd. On an aside to your last sentence...... Something like libbitcoin should be the reference implementation, preferably with a run-time plugin architecture for custom coins, and the community should supply pre-built binaries. Then developers across platforms can get their teeth into applications instead of trying to rip out bits out of another application. An RPC interface isn't good enough for developers. |
|
|
|
Then you are not talking about a decentralized consensus on the finality of transactions, i.e. that can't result in a double-spend.
Dude I have a very deep understanding the of possibilities for Byzantine fault tolerance for the CAP theorem consistency, partition tolerance, and access (liveness) in consensus ordering systems. Putting unordered items into a distributed database has nothing to do with it.
Indeed. However I was responding to Jabbawa about the XOR distance function that Maidsafe uses and the interesting features that arise when applied to bitcoin as a DHT. You then called me a liar "spewing incomprehensible babble". |
|
|
|
Write a white paper, otherwise you are just spewing incomprehensible babble. Invariable those who can't write it down in a whitepaper, are spewing incorrect babble.
No. I have a family to feed so the software that I work on is carefully chosen and purely academic papers to prove to game theorists that something has merit is not even on the radar. Add to that the vehement resistance to anything that changes the status quo away from centralisation and there is little incentive for me to do anything like that. Bitcoin is heading towards being a credit card back end and I pretty much agree and feel the same as TPTB_need_war from one of the links in your previous posts. We are not producing any fundamental breakthrough on the problem of decentralized electronic money. I do not like to work on things that I feel are misdirected and destined for failure in the end. I don't want to get rich by fooling other people (or fooling myself). I'll throw a few ideas and software techniques in for others to run with but until I see that centralisation even being talked about as an issue then I have to spend time on software that feeds my family as I'm not independently wealthy nor part of the paid bitcoin industry. What does game theory have to say about altruism? Nodes can be Sybil attacked. Propagation ordering is not proof nor consensus. Write a whitepaper that explains the Byzantine fault tolerance in your design.
Then you still haven't understood. It is the most unordered propagation you can get and not only from 8 connections, but hundreds. The block chain is still used as the proof-that doesn't change. I am merely talking about a delivery and storage mechanism for the block chain which can provide additional assurances to accelerate the distribution whilst still supplying the network with full node capabilities (without every single block on every disk). However. Thanks for at least asking about it even if you can't be bothered to think it through. It gives me a little more hope that there are people in the community that are still thinking about technical improvements rather than get-rich-quick protocol schemes to directly monetise the blockchain. |
|
|
|
Don't lie. The trust failures have not been incorporated into your lack of analysis of the game theory.
Then you need to think harder about how one fills in the blocks between two checkpoints and what happens if a number of nodes feed you incorrect blocks. Hint: Malicious nodes feeding blocks are no different than orphan blocks. |
|
|
|
|
Show Posts
