Topic: Bitcoin 20MB Fork - page 31. (Read 154787 times)

Ah.  So, the cost of computing the hash of a large block vs. a small one is negligible.  Once the hash of the proposed transaction set is calculated then it is included in the block header.  The block header does not vary in size.  Thank you.  I withdraw my proposal.

The miners are hashing the block header, which is always the same size regardless of how many tx are in a block.

The block header contains a hash of all the transactions, but it doesn't contain the transactions.

So does this mean that we have to promote a "fee war" just to please the miners?

Ouch.  I thought I had at least the slightest clue, e.g. https://docs.google.com/spreadsheets/d/1mOTrqckdetCoRxY5QkVcyQ7Z0gcYIH-Dc0tu7t9f7tw, but please do help me understand.

You obviously don't have the slightest clue about how bitcoin works.

Naturally, as the block reward declines over time my proposal will mean less.  When the block reward is gone then my proposal will mean nothing.  My proposal does scale if/when we increase the maximum block size allowed.

It seems to me that miners could game the current system by deliberately mining minimal block sizes.  They would have a higher hash rate and win more often as compared to miners that try to mine larger blocks forcing them to do more iterations.

With my proposal above that would change, the miners would be motivated to mine the largest blocks possible helping to keep the queue of unconfirmed transactions low.

I had an idea; as I understand it the SHA-256 hash calculations dominate the compute time; essentially we are rewarding miners for their compute time.  In SHA-256, the payload of the first 512 bit iteration is 447 bits.  When more than 447 bits are being hashed additional iterations are required.  A payload of up to 959 bits is handled in two iterations.  1,471 bits are handled in 3 iterations.  In general, 512*N-65 bits are handled in N iterations.  Naturally more iterations take more compute power.  We could consider enhancing Bitcoin by awarding only a portion of the block reward based on the number of iterations needed.  A schedule like the following might make sense;

block size	block reward factor
minimal	0.1
small	0.2
medium	0.4
large	0.8
maximum	1.0

The transaction rate is probably the single most important element defining how the solution will and will not work and thus how well it can be defended.  Right now there is no real attack by those who are potentially threatened by a workable currency solution which is an alternate to those which they've got a stake in so defensibly against attack from that sector is not on most people's radar.

Why there is such confidence in corporations and governments to protect our individual freedoms and liberties given the insults we witness on a continuous basis baffles me.  Especially amonst people who are interested in and drawn to an alternate currency solution.  If/when Bitcoin is more overtly attacked, it's defensibly will be the defining feature dictating it's value.  In monetary terms and otherwise.

I'm a bit disappointed that more people are not framing the problem in this manner.  If Bitcoin survives the gross negligence on the part of those who are self-appointed supposedly to defend it it will be one of the bigger miracles of our modern world.

20 MB Fork is really a good idea or atleast 5 MB
Seeing how just last weekend , the blockchain.info was having some problem pushing TXID's because of the regular Big sized transaction fork's .
They were all around 900+ KB , so yep! 1 MB just won't be enough in some time , may be make it 2-3 MB but we need to increase it

Why should it be the least bit surprising that some thought is devoted to decisions that when made today, may effect the next 20 years?

does anyone know when can we wait this fork? (approx)

Well Bitcoin does have a diverse background of individuals with different education backgrounds and perspectives
We get a large breadth of data opinions and ideas as a result.

It's always going to be difficult to make a go of an 'alt' as an autonomous entity no matter how superior it is.

A universally supported 'balancing' or 'reserve' currency which ties multiple diverse crypto-currencies together would be a powerful and difficult to attack solution.  The robustness against dedicated and well funded attack is ultimately what will give a crypto-currency it's strength in most instances where it would be a thing of value, and 'distribution' is such a key element to that that I usually use the term 'distributed crypto-currency.'  Bitcoin will probably be attacked at the fungibility weak-point through licensing and blacklisting when it is sufficiently consolidated and broadly used to make that attack most likely to succeed.

Back when I was first thinking about this issue I was of the opinion that it was already to late for Bitcoin due to the Satoshi-dice spam and other baggage.  I didn't conceive of the 'two-way peg' of Maxwell's form and envisioned such a backing currency to be kind of a 'plug-in' for alts.  It would be along for the ride, provide some low-resource utility, and expect to be supported by the carriers in a specific POW way.  I called it 'bakcoin'.  'sidechains' seem to be a more workable solution.  A two-way peg is probably a more workable solution than 'plug-in inclusion' even if Bitcoin fails as a workable backing store.

If Bitcoin bloats to the point where it is unfeasible to achieve the kind of distribution that would make it defensible against well situated and funded attackers, something else will probably fill this role.  Bitcoin was not really conceived of as doing this (at least that we know of, though what Satoshi was really thinking in the back of his mind is not something I believe is very knowable) so much superior dedicated design is possible.  At the end of the day, it could end up being better for 'humanity' as it were that Bitcoin goes down the Bitcoin Foundation path (to hell) and a more focused design comes to the fore.  Any 'alternative' to Bitcoin would be free to take it's value starting point from a snapshot of the existing blockchain.  I'm actually quite surprised that some alts have not tried this.  Maybe they have and I'm just not aware of them since I don't pay a lot of attention to the alt world.

edit - sp

by nice and honest i referred to that he wont DDOS other pools or such things. but you are right: we must assume they are greedy (its there right to be...they are a businness)


1BLT does change the fact that bigger blocks raises orphan risk for a miner - thank you: didnt think of that until now.


well thats a point where we disagree and i have the feeling we will never find an agreement about it.
it starts with howto define value... IMHO there are many types of transaction without a big monetary but a huge value for society (or just personnel).

i'd say the only way to make this fair is to not exclude anyone.


in the end people will use what they can use and brings them the biggest benefit.

if there is bitcoin-clone without blocksize, mined by three farms (and nobody except them is running a node) many people will use it. i dont see it as superior or that it deserves to win. its just how people work.

Wow, I have to admit I would have never guessed that such a proposal could have generated such a rift and breadth of opinions.  I'm sorta shocked but then again, the world is a big place.

Reality shows us that a lot of miners will mine at a present loss and speculate on future value.
We should not assume that miners are "nice", or "honest", it's a much safer bet to assume that they are greedy.



If blocks are bigger he won't have a "choice", he'll *have* to create bigger blocks by including whatever comes his way that carries a fee.
The miner won't be held back by block propagation time since it's to become O(1) with the introduction of IBLT.
In other words, the marginal cost of including a transaction in a non-full block will be so low for the miner that includes it, that he will have no economic incentive to not include it.
And by doing so he will push the storage cost to the network as a whole, it's a "privatized gain, socialized cost" situation if you will.



By saying that, you make the implicit statement that someone, somewhere will be able to set a minimum fee, which is not the case, the whole isStandard() band aid is a joke in this regard.
If people can send small transactions with tiny fees to miners, the miners will have an economic incentive to include them as long as the fee is > 0.



The system doesn't need low value actors to increase its overall value, on the contrary, including those creates the risk that the system will crumble under its own weight.



That is incorrect. The only choice miners get is "should I include this transaction in my block?".
The answer is easy, it's a simple comparison of the marginal cost of inclusion versus the attached fee, and with IBLT the marginal cost of inclusion will not be constrained anymore by block size.
In other words, miners will mine huge blocks at everyone's expense because they won't directly bear the actual cost of this inclusion.



If you're able to describe an alternative where some system has all the advantages of bitcoin without none of its drawbacks, then very well, go ahead, I'm listening.
This system will either be inferior, or suffer from the same issues.
And if such a magical system actually exists, then very well, it deserves to win anyway.



You confuse reality and speculation on one hand.
Don't worry about reddit, they'll always have doge to throw at each other.

Can't wait for 20MB blocks as that will give way for huge profits in altcoins with smaller blockchains. The bloatchain will give way to more efficient coins who care about the user being in control.
Insane requirements for bandwidth and storage will just make sure bitcoin won't be adopted by average joe in the end of the day.

I also think 20MB blocks and low fees aren't sustainable once the blockrewards drop and blockchain security will suffer from the 20MB fork. In case we go to 20MB blocks, a few years down the road the same people who are now speaking pro-20mb blocks will speak pro raising the 21 million coins cap for the sake of blockchain security. Fees are needed to secure the blockchain once rewards drop.

Predicting amount of txs in the future and putting btcs' functionality at risk in case these txs don't happen is a moronic move.

I wouldn't characterise Gavin as 'insisting' that everyone gets in to Bitcoin.  I think he is making a proposal for the top item on the Hard Fork WishList.
https://en.bitcoin.it/wiki/Hardfork_Wishlist
It is pretty much his job to 'do something' about it, so I don't fault him at all for making the progress he has so far, my position is just that more progress needs be made before this is complete and done.
It is pretty much my job to peer-review stuff like this, (yours too fellow bitcoiners), to offer counter-proposals, to suggest improvements, etc.

In the http://fr.anco.is/2015/gavineries/ conversation, Gavin poses the question: "...how big is 'big enough...but not too small'?"

Satoshi didn't have the answer to how big is 'big enough' either, so he implemented an expedient solution.  
And...This "how big" remains the most fundamentally interesting question.  There isn't a consensus answer yet.  This is a good indication that there is more work to be done here.  The 1MB limit was implemented by fiat.  Satoshi was running the code and put it in back in v0.3 or so, based on his authority of knowing more about it than anyone else, it stuck.


Some argue that we should do whatever Gavin proposes, because we have to do something now or else tragedy befalls us and game over.   The fear is 'ossification', that Bitcoin has become inflexible.  
The problem with this approach is that opposite argument is also valid:  That Bitcoin is too flexible and can change to the whim of special interests which may destroy it, and that Gavin especially ought not have his way because that means control is centralized which makes it vulnerable.
....so I throw out both of these positions.  They are entirely speculative fear-based.  Most importantly, It should not matter who proposes.  If it is the right solution, it will be adopted.
This is what we do not yet have, the right solution.

What we are left with is an expedient change.  The driving fear that the bad guys will take Bitcoin down because it has not yet expanded to handle more transactions.  It is not entirely unreasonable, but neither is it compelling.  Because of this, I remain content to wait for better.

Why?  I don't like extrapolations, I don't like exponential growth, I think x20 is too much of a jump at once (but I don't know what the 'big enough...but not too small' size is either).  I don't like the inherent pessimism of it, (that we won't get to a good answer to this problem for at least 20+ years).

What I do like about the proposal:
1) It is forward looking, and includes some growth over time.
2) It has a historical basis (Neilson's law)
3) It has an eventual expiration (20 years)

As a proposal, I think it has progressed, but I don't think it is close yet to being an acceptable proposal.

I find expressions of certainty regarding the future as cause for suspicion, but I think you sincerely meant this casually, and are thinking it is "unlikely" rather than "certainly isn't going to happen".
Certainty of the future is typically a 'tell' for an unsubstantiated argument by elevating it apriori, and ought be used cautiously.

Consider whether others may reasonably see an opposite future?  One in which the more expensive transactions, especially those where the higher data size is paid by higher transaction fee, may favor high value transactions (such as an SPV proof may be) over a lower value transaction.

The loser in the block size scarcity is not going to be the SPV proofs so much as the microtransactions because the fee may swamp the value of the transaction.  Things like pay per click adverts and changetip cashouts would have to raise the minimum transfer.

As much as I would like Bitcoin to be more useful, it is not worth introducing new unchecked risks.
The block size expansion does increase the trust placed in miners, there is not any getting around this.  Of all the places to put trust, this is likely the best, as it already exists.  Even so it MUST be done with extraordinary care.  Increasing necessary trust arrangements with Bitcoin is not to be done at all if possible, and only if necessary then done as minimally as possible.