Subject: Block generation speed not so important long-term (and the analogy between future bitcoin payment system and today's central banks, AND what is different (=better) with bitcoin!)
[bitcoin, future vision, long-term, online wallet services, banks, wallet accounts, retailers, stores, payments, block chain size, block size, transaction fee, tx fee, scalability, confirmation time, fiat, central bank, analogy]
Hello again,
it has been talked a lot by alleged mathematicians or psycho
pathslogists about whether the number of zeros before or after the decimal dot and the value of the nominal cap limit of a coin is important or not, and if yes, in which direction. I am tired of that discussion, there is no more substance added to it.
So today I am writing about something completely different, which is neither "mathematical" nor "psychological", but rather "economical" and "technological". I have thought about how the future may look like, taking into account technical constraints of bitcoin (or actually "anycoin!") and market developments:
Bitcoin block size is limited. There is currently a limit of 250 kBytes ("soft-limit"), and the final hard-limit of the protocol is 1 MByte. There is some speculation whether that limit will be eventually lifted, but this is very unlikely, because this would not only mean a protocol change, it would also accelerate the growth of blockchain size without fundamentally changing anything, because the next limit would be reached soon, and you cannot increase the limit indefinitely.
Today the block size in the blockchain has already increase a lot recently, it is now at around 210 kBytes average, after about 50 kBytes around 1 year ago,
as you can see here. So we are going to hit the 1 MByte in a not very distant future.
This means that the number of transactions that can be carried out over the bitcoin network per every 10 minutes is limited. What will happen (and actually already has started happening) is that transaction fees will start increasing in free competition. If your tx fee is too low, your bitcoin transmission will not make it into the blockchain, because miners will prefer other transactions with higher fees.
This applies for transactions that actually get carried out over the "real" bitcoin network.
Moreover, we have online wallet service providers like "Mt.Gox", "blockchain.info/wallet" and many others. If you make a transfer from your online wallet account to another account of the SAME wallet provider, you do not need to transfer bitcoins via the blockchain! Instead, you just need to know the "address" (username or whatever) of the recipient. It is like in banking today: If you make a wire transfer to another account of the SAME private bank, this bank just changes some entries in its ledgers, but the central bank does not get involved. However, if you transfer the money to an account of another bank, then your bank has to actually transfer "real money" (=central bank money) to the receiving bank. That's how banking works all over the world today. So what is the "central bank" in fiat world, that is the bitcoin network/protocol/infrastructure in the bitcoin world.
So what does this mean in practice for the normal consumer that wants to buy his bread, milk, or new trousers at the retailer around the corner in bitcoin?
Well, the retailer has the following signs sticking at the entry door:
VISA ACCEPTED MASTERCARD ACCEPTED AMEX ACCEPTED Mt.GOX ACCEPTED
BLOCKCHAIN.INFO/WALLET ACCEPTED
XYZ WALLET SERVICE ACCEPTED
DIRECT BITCOIN TRANSFER ACCEPTED
So you as the customer have the choice. Maybe you do not have an account at "MT.GOX" or "XYZ", but you do have a "BLOCKCHAIN.INFO/WALLET" account (and the corresponding app on your smartphone), and you may also own some "real" coins accessible via your smartphone app "bitcoinspinner" or "android wallet".
When you go to the checkout to pay your bread and the liter of fresh milk, the cashier asks you: "How would you like to pay?"
You say: "blockchain.info, please".
Alright, so you pay with your blockchain.info app to the retailer's "local" blockchain.info-address, and the payment will be confirmed almost instantly by the blockchain.info server at the retailer's blockchain.info wallet account. So you can walk away with your bread and milk quickly.
This transaction was free of tx fees, because you did not pay via the bitcoin network. You could have payed via the bitcoin network ("DIRECT BITCOIN TRANSFER"), but then you would have been required to wait, or you would have had to buy a voucher of this store in advance and would have paid with the voucher now. Another disadvantage of the direct bitcoin transfer method is of course that you would have had to pay transaction fees to get your transaction into the network in the first place. By that time (depending on how long we are looking into the future), the tx fee might take a considerable part of the purchase, if you just bought a small bread.
So the big idea is that all the retailers will have (free) accounts with all the major wallet service providers, such as to make sure that every possible customer will be accommodated.
This is my vision of the future, and I think it is not just a possibility, but a bare necessity that we have exactly this kind of development of the bitcoin (or any "non-niche-coin") ecosystem, given that bitcoin succeeds. And this is not at all enforced by the 10 minutes long transaction times of the Bitcoin network, but by the pure scalability problems, as explained in the beginning (block chain size, block size limit and resulting increase of tx fees).
Now some Q&A, to prove wrong the statement "but I don't like this, then I can equally well stay with today's fiat money":
Q: But this means that I, as a customer, have to have my bitcoins in my XYZ wallet provider's account (or another wallet provider) if I want to pay in retail stores. I don't like this. I do not trust any wallet provider. I want to keep my bitcoins locally, in my own wallet, with my own private keys.
A: Don't worry. You do not need to (and actually should not) move ALL your bitcoin savings to your XYZ provider's account. Instead, you are advised to move just as many bitcoins as necessary to your XYZ wallet as you need for your next shoppings (maybe for 1 week or 1 month). Then, if XYZ gets hacked (or claims so), your loss is limited - it is just as if you loose your wallet with your cash today (or rather... 20 years ago, because today you have cards, not cash, in your wallet
). You may also carry some "true bitcoins" with you in your smartphone's "bitcoin wallet (or bitcoinspinner)" app, in addition. Then, if you find out during your Saturday's shopping tour that you run out of bitcoins on your XYZ wallet's account, you can just recharge your XYZ account from your bitcoin spinner during your coffee break (of course with some tx fees), just like you had to go to an ATM in "pre-credit-card-accepted" times to continue your shopping tour.
Q: In this "vision of the future", it looks to me as if the wallet service providers act like banks: The customers open a (bitcoin-)bank account called "online wallet" and deposit (some of) their bitcoins, and then payments are carried out via these bank accounts. Isn't this a mere copy of our current banking system?
A: There are indeed parallels, exactly as you describe it in your question. And you can take the comparison even further: Like the money that you have on your bank account today, also the bitcoins that you have on your wallet provider's account is not actually true money, but just a liability of the bank/wallet provider to you, the customer. True money is only "central bank money" that the private banks have in their accounts with the central bank, or money bills and coins. In the bitcoin analogy, one owns true money when one is in sole control of a bitcoin private key. A major difference between today's money system and the bitcoin money system is this: In today's money system, private persons or companies cannot open electronic "true money accounts" at the central bank. This is a privilege of the private banks. If a private person or a company wants to store some true money, this can only be accomplished by hoarding bills or coins. This entails some risks of loss or theft, which is the main reason why most people still have their money at the banks today (despite almost zero interest rates). Instead, in the bitcoin world, "true electronic money" is accessible by every market participant ("banks=wallet services", companies, private persons) alike, they just need to own a private key and have some bitcoins sent to the corresponding address. This makes safe storage easy, simple, safe and cost-free, and nobody needs to fear to loose a big fraction of money due to a bank run or bankruptcy of a bank.
Q: I see, but I still don't like this future, because these "wallet services" could run a fractional reserve system, just like today's private banks (legally) do, thereby inflating the amounts of bitcoins.
A: True, the wallet providers could run a fractional reserve system. But there are two big differences to today's fiat money system:
(1) First, they would have to be much more cautious than today's banks, because they cannot be bailed out by the central banks (who can "print" as much "true"(!) money as they need) to keep up liquidity. Instead, in the bitcoin world, the amount of "true" money is limited to 21 Million, and that's it. So you cannot create money bubbles by printing more "true" money. Instead, the "bank" (=wallet service) would crash in due time, with some (limited) damage to the customers (which is why they should keep their big savings in their own private keys, not in online wallet accounts), but after such a thunderstorm the air is clean again, without an ever-growing money bubble poisoning the economy.
(2) Secondly, there is a mechanism by which any wallet service provider can reasonably convincingly prove that he is NOT running a fractional reserve system.
I have described this method here (note at the time of writing this the whitepaper in version 0.1 is published. I soon want to publish another whitepaper with some additional improvements). Of course, no wallet service provider can be obliged to do this, but if the community is sufficiently aware of this problem and the method to prove it, it will hopefully become a competitive advantage.
Q: What would be the business model of these wallet services ("bitcoin banks")?
A: There are many possibilities. One idea is that they act as exchange services (like MtGox) and payment processors (like bitpay) and offer the wallet service as additional free(!) service to both businesses and private persons.
Conclusion:After all, I pointed out that on the long term, if a crypto-coin succeeds and gets used at large scale, the main "business" is going to take place outside the block chain, with bitcoin accounts of wallet service providers. This evolution is unavoidable due to scalability issues (blockchain size and amount of transactions per time) that any crypto-coin would face. Also, the tx fees of the bitcoin network will grow considerably compared to today, due to the competition of the transactions trying to get into the block chain.
With "main business" I am mainly referring to purchases at retail stores or online stores, but also micro-payments or small donations over the internet, i.e. all payments that are small relative to the tx fees.
Bitcoin payments on the blockchain will still be used for example to load or withdraw funds to/from an online wallet, or to make greater transactions were tx fees do not matter. Also payments between wallet providers (that may happen once per day, e.g. like payments with central bank money between private banks today) will still use the block chain of course.
Given this, it becomes evident that confirmation time (i.e. block generation speed) is not very relevant for the successful long-term development of a crypto-coin, and for sure the 10 minutes confirmation time of bitcoin is not too long to be a problem. 10 minutes is just as good as 1 or 2 minutes or 10 seconds, because the main business is anyway going to take place outside the block chain. Note that anyway also a 1 minute average(!) confirmation time is too short to be practical for retail stores.
A 10 minute confirmation time, like for bitcoin, combined with a reasonable block size limit, even gives an incentive to the community to develop wallet services, as outlined above, in due time, especially to circumvent the pain of increasing tx fees. Such wallet services will first predominantly be used online, as usual, but once such services exist, they are likely to ACCELERATE the more wide-spread introduction of bitcoin payments in retail stores. So after all, the slightly "longer-than-actually-necessary" confirmation time (10 min), combined with a "tighter-than-perhaps-needed" block size limit (1 MByte), could even turn out to be an accelerator of the evolution of bitcoin infrastructure and wide-spread bitcoin adoption.
I would not be surprised if Satoshi has carefully thought about all this well in advance and designed the system parameters accordingly.