Bitcoin Forum

I had an idea for a way to decentralize mining. It was quite odd actually. I had been sleeping and woke up in the middle of the night. Suddenly the light bulb went off in my head. People have taken different approaches to prevent big megamines but it really hasn't worked (different algorithms). That's because the underlying system is a bit flawed in preventing centralization.

Now as I understand it with most coins, a new block is generally generated every 10 minutes.

My question was, why? Does it have to be this way?

So, I got to thinking. This short block time creates a lot of the problems with new coins. People claiming there was a premine etc. Also, it doesn't combat centralization. There's nothing to encourage more network nodes. We want more nodes. That's my opinion.

It also makes people reluctant to get in on new coin after a lot of the early coins have been minted.

So, I got to thinking. Could you tie the amount of time between blocks to the size of the network (total hashrate)? No, this wouldn't work because you could still have big megafarms plugging into the network.

So, my second idea was nodes...tie the amount of time between blocks to the amount of active full nodes. If you had one miner on the network essentially a block would never be generated. Once you had two, maybe a block would be generated every 24 hours (just an example). If you had a million full nodes on the network maybe you could eventually have the time between blocks be like 10 minutes as most standard coins are right now. This would make the money generation rate more dynamic. It also would generate money based on active users.

Could this be possible? If so, anyone interested in working on this concept?

We could really make a decentralized coin this way where many people are encouraged to run nodes.

Long block times are awful because transactions will not be confirmed for ages. Even 10 mintues is way too slow for today's standards.

And your solution doesn't incentivize people to start their nodes. They'll just rather ignore the coin and let it die because of the blocktimes.
Maybe you'd want to change the block reward and lower it if the node count is low and increase it if the node count is high.
The problem with that is that nodes can be attacked and knocked off the network easily.

There's also a difference between miners and nodes because of pools. If pool mining were impossible then miners would all have to run their own nodes and mine to that.
But with pools miners doesn't have to run nodes, just transfer funds directly from the pool to anywhere.

Spreadcoin was created to solve that problem by making poolmining impossible. While it didn't succeed (there's at least one pool) but people aren't poolmining it.

It's a bit different but there are also coins with masternodes. Some of them have thousands of them running actively.

Ok, but why can't we make the block rewards separate and confirmations there own separate thing?

Ooo, I also had another thought.

How are transaction fees distributed now?

I think it be nice if every time transactions were confirmed the transaction was redistributed evenly among all active nodes. Could you do that?

Where do all the transaction fees go now?

It would be easier to distribute individual segments of blocks to different miners running.
Alowing for shares between miners. Also adding a cap to the amount og GHz connected could also work possibly.
Or for every user to mine the coin at the same time and geting a very small reward for doing so.
The first person to crack a block gets the greatest amount and everyone else who cracks the block gets a small amount too, helping with distribution?

You just reinvented Proof of Stake.   :)

BTC & Peercoin use a 10 minute Block time.
LTC uses a 2.5 minute Block time.

ZEIT Block Time is 30 seconds.
in the Beginning was PoW/PoS now PoS only.

Supercoin has a 20 second block time.

We earn interest , each Block with coins in it can generate a new block when it stakes.
(That is PoS incentive to run a node.)

The only reason for the longer block times was to keep the Blockchain rate of growth controlled.
The more blocks generated the more entries in the ledger and the more space it takes up.

Instead of making the Blocks bigger for BTC so it could hold more,
they could have hard forked and went to a 5 minute block time to double their capacity or even a 2.5 minutes to quadruple their capacity.

 8)


FYI:
I will tell you , I luv sending ZEIT and having it arrive and confirmed in only 2 minutes.
Where BTC  & LTC are down right slow in comparison.
However if your wallet has been offline for a few days, the longer block time coins will sync with the network faster , due to having to download less blocks.

That's what we do....except we don't have blocks.

Our consensus algorithm EVEI doubles up as a kind of "Record of Service", so that nodes can be rewarded for the work and service they have provided to the network and is accurate to within 10-15%, thus the node incentivization problem is solved.

Quite a few smart minds have looked at this, but we're the first to solve it (it's been solved for about 18 months in fact)

Exactly, currently BTS has fastest sending and capacity btw. ETH wants to achieve it soon.. BTC is in trouble for now.

Your idea with block time depending on number of active nodes is interesting. I must think a bit about it to come to a real conclusion if the idea is sound or not and how it can be accomplished.

Obviously, there should be a kind of "minimum block time", for example, a 30 minute interval, and then a formula could be applied how much the block time decreases with each new active node.

A problem I see, at first glance, is that the inflation will be somewhat impredictable and can get pretty high. That's also the problem Timekoin (the most similar system I know) has (see http://timekoin.org for more info). But I like their approach of rewarding active nodes permanently.


That would mean thousands or even millions of transactions per block, because you would have to send everyone its "share". It would also mean a lot of dust and very small amounts.


Timekoin also has a similar system, and it's a little bit older than eMunie (2012). There, active nodes get elected and then they receive rewards each block until they go offline for more than 48 hours approximately. I'm not totally convinced of the Timekoin algorithm though, because a 51%-attack looks pretty easy there.

The act of being able to create deterministic block producers at all signifies you *don't* actually have a decentralized currency in the first place.  Larimer was kinda sorta upfront about this by the fact that he called Bitshares a "decentralized corporation".  In reality, it's more like a "distributed" corporation and not decentralized.

In other words, Bitshares and anything you can create using proof of stake, recursive systems is not in the same category as Bitcoin at all.  They're completely unrelated systems.  Bitcoin is an actual decentralized currency, and nothing you can create with proof of stake is.

As for the Fuserleer system, creating decentralized systems is the same thing as creating a multiplayer video game and trying to proof it against cheating with the knowledge that anyone can modify the client.  He's always been vague in how the thing is supposed to work, and I feel he's creating the equivalent of a traditional video game that people will implode by modifying the client since no outside resources are relied upon for convergence.  He's going to say I'm wrong, but it's his job to actually prove it's secure in a white paper, and I don't think he can do that due to the system being the equivalent of a giant Rube Goldberg machine.

Fuserleer is pretty smart, but I'm definitely an Emunie skeptic o_O

I'm pretty sure there are cryptos out there with 1s blocks.  And we de-bunked the whole 100,000 tps hype quite some time ago :)


Right, determinism itself isn't a totally bad thing though, so long as the signals which are used for said determination have a strong element of random entropy that can not be easily monopolized.

We use determinism a lot, especially in the consensus, but the random element is large enough that the output can not be predicted beforehand.


I'd rather be vague and frustrate people than jump the gun and make a statement which can not be retracted later. 

Just taking on one part of a crypto platform is hard work, I'm trying to solve scalability, efficiency, currency stability, infrastructure issues and more in one hit.  I'm not settling for just moderately improving in these areas either, I want a leap forward.

I'm in no rush, I can afford to take my time and iron out all the kinks.  For the most part it's almost kink free now, so hopefully I can provide some satisfaction soon enough to make up for the frustration.


*faints*


Ahh back to reality :)

I honestly think such an implementation (where nodes are inversely proportional to block times) would have the opposite effect of what you are suggesting.

Regular users and hodlers would be incentivized not to host a node. Longer block times = lower inflation = higher intrinsic value per unit of currency held by users.

Big miners will host more nodes to create more blocks and thus profit further.

The end result is an even bigger divide between people and miners, and miners having even more power and say in the crypto. Such a specification would make an environment worse than what we already have.

Hmm, interesting. Are you talking about the beta coin in your sig?

Yes, I need to think it through a little more. Maybe run some math through my head and think of some basic relationships like Moore's law etc.

Maybe you could have the block time related to the nodes and the amount of coins received with each block related the hashing power of the network.

That way if you had a ton of miners operating nodes but no cpu power block rewards will be quick but small. For hashing power you could do an exponential relationship of power to number of coins. As power increases the coins increase but as the cpu power gets exceedingly high you get diminishing returns in coin increases do the logarithmic relationship.

I think it would make for a very dynamic network. It would at least be a very interesting experiment. What can it hurt to try? We could also make the transaction volume high right off the bat.

Edit: Okay, I figured out how I would want to do the block rewards. The computing power. It would be the log base 10 of computing power in bytes. So, at 1 byte coin reward would be 0. At 1 KB the block reward would be 3 and so on and so on.

Now for the block timing, I need to think that through a little more. If anyone is interested and wants to help let me know.

really like the idea it really have to change the decentralization system