Quote from: AlexGR on March 11, 2016, 12:13:56 PM

Quote from: Andre# on March 11, 2016, 10:06:45 AM

Suppose a bus company starts running a service in 2010. At first, you'd only see one or two people in the bus. As the bus service becomes more known, more people travel on the bus. Nowadays, busses are often completely occupied. It's common that people are left at the bus stops having to wait for the next one, esp. if they bought the cheapest tickets.

The difference between buses and blockchains is that buses can only take onboard real passengers, not ...fake ones.

Blockchains can be burdened with millions of transactions just because one guy, that didn't even want to make one legitimate transactions, wants to have fun.

Quote

I can imagine that a flood of tx without fee can be considered spam. But nowadays, only mining pools use zero fee tx to payout their miners. There are hardly any zero fee tx issued otherwise. (800 in the past 24 hours, source: https://bitcoinfees.21.co/ ) But if a tx carries a fee, how to decide if it's spam or not? What is your definition of spam?

Someone "pays" for 1mb of spam with 1 satoshi per byte. Thus for occupying 1mb of space, he only needs 1 million satoshi = 0.01 btc = 4.2$.

He can actually occupy the entire's day blockspace of 144mb by "paying" 604.8$.

Even at 2mb, he can buy a block at 8$ and a day's worth of blocks at 1200$.

Does this mean that he is not a spammer because he "paid" a ridiculously low amount of money?

So the definition of spam is something that is bought cheap? The water coming out of my tap is cheap, so if I use that water I'm spamming the water supply?

Currently, 1 MB blocks can be bought for about 0.5 BTC, or $200. If tx are too cheap, the price should be raised to the point it becomes economical. Instead of putting a limit on supply.

Quote

Quote

Fact: the number if tx per day is close to the limit of 250,000. We recently touched that twice. (sources: https://blockchain.info/charts/n-transactions?timespan=2year&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address= , http://www.coindesk.com/data/bitcoin-daily-transactions/ )

We crossed the 20,000 tx/day in June 2012
We crossed the 50,000 tx/day in August 2013
We crossed the 100,000 tx/day in March 2015
We crossed the 200,000 tx/day in January 2016

And the hard limit is a little over 250,000...

So, do you think 250,000 tx/day is sufficient for Bitcoin to be successful? Do you think 400,000 tx/day is sufficient by the end of this year, when SW is rolled out and most other software is updated to take advantage of it? Do you think 400,000 tx/day will be enough until LN comes into existence?

Compare these:

Number of transactions: https://blockchain.info/charts/n-transactions

vs

Number of Transactions Excluding Chains Longer Than 10: https://blockchain.info/charts/n-transactions-excluding-chains-longer-than-10
Number of Transactions Excluding Chains Longer Than 100: https://blockchain.info/charts/n-transactions-excluding-chains-longer-than-100
Number of Transactions Excluding Chains Longer Than 1000: https://blockchain.info/charts/n-transactions-excluding-chains-longer-than-1000
Number of Transactions Excluding Chains Longer Than 10,000: https://blockchain.info/charts/n-transactions-excluding-chains-longer-than-10000

(long chains = moving money to next address => to next address => to next address => to next address => ...)

From these graphs I conclude that about 2/3rd of all tx are part of chains longer than 10. So what? I suspect those who do these tx have a reason for it, and are willing to pay for it.

The fundamental problem seems to be that the true cost of a tx is not covered by the fee. That's why people who do long chain transactions are not discouraged by the costs of doing so. If this is a threat to Bitcoin, the price of tx should be raised (a minimum fee that truly covers the tx cost), or the efficiency should be increased (bigger blocks, LN).

Question is, what should the minimum fee be at a given block size limit to have enough headroom for a consistent user experience? Recycling an arbitrary limit as a capacity quota is not a good answer.