Topic: Gavin Andresen Proposes Bitcoin Hard Fork to Address Network Scalability - page 3. (Read 18415 times)

But this won't solve the problem over time.  See the response by kkaskal in this thread: https://bitcointalksearch.org/topic/title-near-zero-bitcoin-transaction-fees-cannot-last-forever-668704

And the graphic at URL in this thread:

https://ip.bitcointalk.org/?u=http%3A%2F%2Fi.imgur.com%2FMCWO0tH.png&t=545&c=8_MsRYuvimAWfQ" alt="" border="0">

Note block reward is decreasing over time, a well known problem since the number of bitcoins is fixed and it gets progressively harder to mine bitcoins over time, all things being equal.

One of the main reasons these things haven't been implemented before is that Bitcoin as a whole was still in a massive alpha stage, having more transactions per second simply wasn't needed up until now. And now especially so that we can continue to grow.

Commercial profit driven companies attempting centralized solutions (like blockstream) to problems can be solved with decentralized and open solutions are honestly very terrifying, and I hope that technological advances on Bitcoin deploy fast enough to prevent any side-chaining or centralized solution having to take its place so we can continue this open concept revolution for everyone to continue spreading. 

What is also important is sidechains are useless if you want to receive your remainder of tokens back in bitcoins. All this will do is switch value away from an open decentralized solution to a closed centralized "solution".

Nice of you to correct me, and maybe ledger isn't the right word at all considering that most of their transactions are credit based, but do you really expect me to believe that Visa contacts JP Morgan Chase to verify credit hurdreds of thousands of times per day?  I'm simply saying that Visa is keeping track of a lot of data in order to process all those transactions.  It may be pruned regularly since the banks have to provide statements and whatnot, but some transactions take a few days to clear, so they certainly have to track it that whole time.  Regardless, none that changes the fact that they process and track a lot of data.  Moreover, my point that storage is the cheapest part of a miner's arsenal still stands, and as long as miners are paying the kind of money they currently pay for ASICs, storage is not going to concern them.  On the other hand, income obviously will, and IMO, when the block reward halvings get relevant, more transactions are more likely to keep bitcoin secure and competitive than higher fees per transaction and users that go to sidechains to avoid paying the fees (and that's even assuming someone as genius as Satoshi comes up with a good trust-free way to make sidechains serve this purpose instead of all the purposes sidechains were really meant to serve when the concpet came about).

Clearly a simple but good move forwards to massive bitcoin adoption, hehe.

Sorry, I know that for the last few posts I've been talking about side chains rather than the proposal.  

This is relevant because I'm trying to demonstrate that unless someone makes a fairly profound cryptographic invention, side chains simply are not a trustless solution.  And in the long term nobody can be a point of trust without offering a point of failure and point of control to all the scammers, cons, extortionists, terrorists, dictators, governments, etc in the world.

We need this fork.  We don't have a viable alternative that can retain Bitcoin's permissionless nature.  

I've heard all this kind of things ad-nauseam for the 3+ years that I've been following things.  Simultaneously I've seen both the hardware and network improvements which have been made realistically available over that time period, and also the struggles Bitcoin has had just to make the original 7 TPS number kinda work.  Many of them were not pretty.  Some of these rosy predictions about how technology will swoop in to save the day (and let us go from 7 TPS to 200,000 TPS which is enthusiastic even for your type) would be more convincing if there were a few more signs of it actually occurring.

This only works if you are eventually cashing out the same number of tokens you cashed in for.   The minute someone actually pays someone on the side chain, and doesn't eventually get back an equal number of sidechain coins, the people they paid on the side chain have no way to get their money back to the main chain without the payer's continued cooperation.

Suppose Alice uses a Bitcoin to get 1000 mBTC tokens.  She ripens them and pays Bob 750 of them on the sidechain, then cashes out her 250 remaining mBTC tokens back onto the main blockchain, sells all her BTC  at an exchange, and takes her client offline permanently.  How does Bob ever cash his 750 tokens out without her cooperation?  Alice hasn't even stolen anything here, but she's still just being a heartless bitch. Maybe Bob is her ex-husband or something.

So essentially you still have the "Trent" problem, except that now "Trent" is whoever happened to make the original BTC-locking transaction.  Nobody can cash out their coins unless Trent is cooperating, and Trent won't cooperate when under duress or extortion.  Nor will Trent cooperate if Trent is Alice and just doesn't give a crap whether Bob lives or dies.

I've thought of that, but it doesn't work unless the sidechain coins come in non-divisible (and more to the point, non-unitable) denominations.   Otherwise the 2nd sig becomes "problematic" when facing a sidechain coin that represents a microscopic interest in each of thousands of BTC-locking transactions on the main chain -- which happens, on average, after about ten spends if the sidechain coins can be split and recombined like Bitcoin.

And you have to have Trent holding the 2nd sig of the multisignature.  The minute someone serves Trent with a photo of his wife and kids with guns to their heads, nobody can redeem their coins off the sidechain without the extortionist's approval.

Visa does not actually keep a ledger. If you wanted to use a Bitcoin analogy, you would say that Visa acts as a node (that does not keep a mempool of unspent inputs) and will only "accept" a transaction if the issuing bank confirms there is enough unspent inputs (credit line available).

Actually combining blockchain pruning (i.e, the "main" blockchain only storing UNSPEND addresses), and some of the new emerging technologies like ed25519 that even on an old 2010 CPU we're already capable of 80,000 verifications per second. Seeing how we have a HUGE HUGE way to go before we actually reaching that amount of transactions per second, when we actually do reach it would likely be way post 2020, where we will have even faster hardware and likely better algorithms, which should technically be able to sustain the guidelines I mentioned.

Not to mention that most people would be running light clients anyway around that time, just as described in satoshi's original paper.

In the end there is no reason not to have everything running on the main blockchain since technological difficulties can be overcome.

Pretty sure satoshi talked about light clients for the masses.  Also pretty sure VISA keeps a ledger.  Are you suggesting that "clued in guys" have insider knowledge that storage technology isn't going to continue to grow and get cheaper?  Do you also think "clued in guys" are writing solutions that pay people to mine things other than bitcoin are doing so completely philanthropically and not out to make a buck?  Is there any real evidence that sidechains aren't just going to be the new altcoin?

ETA: Considering the cost of ASIC mining equipment, storage technology would actually have to start shrinking and increasing in cost to the point that individuals couldn't afford computers before it would begin to matter to miners.

As usual, your theory sounds great.  However, I've seen nothing that would convince me the block size shouldn't be increased.  That having been said...Blockchain pruning is very important.  Blockchain pruning and blocksize increases are not mutually exclusive.  I hope blockchain pruning is being worked on, and optimally, blockchain pruning would be included in the same fork as blocksize increases.  Data for sidechain transactions has to be stored and secured as well.  In my mind, using one super-large super-secure network is better than having to move between a lot of competitive networks when you are talking about decentralized systems like Bitcoin.  Yes I'm glad most local retailers accept Visa AND MasterCard, but I'd be better off if they didn't an I wasn't paying the fees.  To that end, if they start accepting bitcoin and offering a discount for its use (which admittedly may only be as likely as businesses offering discounts for cash, which are rare), then I would expect both discount offers and acceptance to decrease with each additional sidechain.  IOW, I would expect the same nonsensical multi-tiered payment processor plans where you pay more to accept SC1 in addition to BTC and even more to accept each sidechain beyond that.  With that kind of potential outcomes, I believe it is much better for bitcoin to adapt as necessary and even expected from the start than for it to be treated as an untouchable behemoth base that must be worked from/around.  However, back to the first quote, I also believe that most countries where bitcoin will be useable in the beginning are supposed to be controlled by their citizens and that, as such, bitcoin should not need to be modified (or connected to sidechains) for regulatory reasons to begin with (nevermind my previous comment that most regulations control corporate entities and bitcoin is already more visible than fiat, so certainly shouldn't need more control than fiat [which can't truly be controlled anytime it is in the hands of citizens]).  Nonetheless, to the suggestion that bitcoin or a sidechain could need changed (or created) for regulatory reasons, I say it is incredibly unfortunate that fear in uninformed citizens and money from large corporations seem to be the only things driving the law most of the time.

That should make it pretty clear why I (and quite a few others) don't believe that Bitcoin in anything remotely resembling it's original and described implementation (distributed persistent decentralized ledger) is applicable as a solution to the described problem.  Bitcoin won't scale just because people would think it's cool if it did.  A handful of large organizations probably could run Bitcoin at these rates but realistically that's about it, and at that point they may as well do away with the blockchain itself and make it a real-time system like VISA.  That's why clued in folks like the blockstrream guys are looking around for a solution which will let Bitcoin scale to these levels without abandoning the very desirable aspects of the solution that it has today.

Transaction fees only exists for 2 reasons currently;

1. To prevent massive blockchain spam
2. To keep the miners going after block reward is either extremely low or non existent

If no more miners would be build it would actually be more profitable for current miners because the difficulty would stop increasing exponentially, current miners would be celebrating on the streets since they will keep making the same amount of bitcoins per month.

On the counterpart, old mining equipment simply goes into retirement when they become obsolete (too slow, too much power draw, etc)

I don't see the average consumer dealing with sidechains and everything. If we remove the limit of 7tps soon, we might be in time to prevent massive transaction fees when the next price/transaction spike happens. In the end Bitcoin is advertised as a low fee system, if each transaction starts costing dollars personally I would loose interest.

Also creating sidechains would destroy Bitcoin when the block reward gets extremely low or when we stop having block rewards, since miners on the main chain would either stop mining since the transaction fee is so low (making the network extremely insecure), or make transactions on the main chain insupportably expensive to maintain their mining.

In order to prevent high transaction fees we need to increase blocksize and remove the 7tps limit. It appears that the majority of community is in "pro" of this so it should be done quickly to prevent the inevitable high transaction fee explosion. When transaction fees become important hopefully miners will make their fees by high volume of transactions (all at a low fee).


Example:

Bitcoin high frequency trading:   

47,000 tps (around the peak rate of VISA) = 28,200,000 transactions per ~10 minute block

Fee of 0.000000017 = 5 bitcoins per block transactions reward for miners. Easily sustaining the network at a high price. If Bitcoin is going to be a global cash system we'd probably have hundreds of thousands of transactions per second. But having more transactions per second than VISA is a good initial.

High transaction fees:

7 tps limit:

5 bitcoins (transaction reward) / 4200 (tranactions per block) = 0.001190 transaction fee. Way too high for the consumer paying his hamburger at a restaurant.

=====================

Combining pruning + near no limit of transactions = low transactions fee.

Why artificially limit a open concept market? The market will take care of the fee, we shouldn't impose limits on it except to prevent massive blockchain spam or otherwise malicious activities.

There needs to be a reasonable supply/demand curve in order to get fees to be a voluntary thing and allow the market to work (though the inflation-based block reward currently fucks up the market royally.)

One cool thing about sidechains (and probably other constructs as well) is that they create the potential to aggregate native Bitcoin transactions.  This means that very significant transaction fees could be paid to Bitcoin infrastructure providers yet individual users would hardly feel it because they share the cost.

Indeed, even if no more mining rigs were built between now and 2040-ish it would require very large transaction fees just to keep the existing rigs powered up unless BTC valuations expand by a bunch.  And if the miners cannot make a dime helping Bitcoin then we better hope and pray that they cannot do so by harming it either.  Of course 'hope and prayer' are often fairly ineffective engineering strategies.

I think raising the fees for scalability would be too much. For a fee of 0.02, imagine if 1 Bitcoin was worth $1000. That would result in a fee of $20 for any transaction, regardless of how much is being sent! That's not very ideal. I think what can be done is to support a larger number transactions per block. More transactions processed = more fees collected.

After thinking about it for a few minutes the one solution which came to mind was something along the lines of what follows, and doesn't even include any fancy stuff though the idea of figuring out how to exchange private keys themselves (possible 'BTC fractions' outlined below) has always appealed to me for a few reasons:

I choose to fund a sidechain with 1BTC.  In doing so, the sidechain gives me 1M unripened tokens.  Only I can ripen them and make them usable.

The sidechain keeps kind of a 'till' filled with whatever fractions of BTC they have handy, but they are locked up and it requires X number of ripened tokens to unlock.  I can trade my ripened tokens around with other users within the system, and if I want to cash out (into Bitcoin) I can dedicate my ripened tokens to the task.

Possibly the sidechain in order to perform optimizations and what-not would request that I give them some ripened tokens which they might use to thaw some BTC.  Possibly just to get their till organized and also maybe sometimes to balances between themselves and other sidechains.  This I would happily do and of course accept some unripened tokens right away as replacements.  Naturally this would be effortless and automatic on my part though I might tweak some knobs once it a while.

Actually, come to think of it, maybe I have full and unique control to unlock specific BTC fractions using my own ripened tokens and the balancing is just me giving up and taking control of BTC fractions as need be to maintain some approximate valuation which I have in the system.

It must be remembered that the source code which performs these operations is open and written to be secure, and part of my decision to use a particular sidechain would be associated with how many eyes had looked things over and their history of non-failure.  Even the system cannot, say, ripen my tokens (unless someone slips some evil code in.)  This is not unlike the methodology which protects Bitcoin proper users.

If you read Gavin Andresen's blog and posts, you will see that in the past he has made the logical suggestion that the users of BTC pay for network scalability by paying more in bitcoin that the current 'voluntary' contribution of 0.1 mBTC.  This is because with increased scalability, the transaction fee must rise to as high as 0.02 BTC, which is at today's prices about $8 USD (and of course if BTC goes higher, this fee will be higher as well), since miners will no longer subsidize low fees (a paper was done showing this, see: http://www.coindesk.com/new-study-low-bitcoin-transaction-fees-unsustainable/   )

As BTC transaction fees rise, G.A. has said, users can either pay more by selling their anonymity to a marketer like Google (a blog post said this, I don't recall where it is), or, simply pay the extra transaction fee money.  

IMO what will happen with BTC as transaction fees rise will be that it becomes more like PayPal or Western Union.

So essentially an SC client will issue a 2 of 2 multi-signature transaction in which you deposit bitcoins. Your client will broadcast the transaction to the SC miners who will send you the SC address to send the bitcoins. It will then automatically send you an SC asset (or coin). Your SC key is then an encrypted version of your half of the bitcoin key. In order to decrypt your key you must transfer the SC into a  conversion account this time purchased from a miner, and give it your Bitcoin address. The algo then automatically broadcasts the 2 of 2 transaction to your Bitcoin address. SC mining would convert coins back and forth by encrypting and decrypting the Bitcoin addresses with SC  tokens.