Bitcoin Forum

I was wondering how come a block with no transaction other than mining reward is a valid block. Is there a technical advantage?

Example:
http://blockexplorer.com/block/000000000000042f7c236a243e2d84bec12210dd2dd5abbeaaf55024898ac70f (http://blockexplorer.com/block/000000000000042f7c236a243e2d84bec12210dd2dd5abbeaaf55024898ac70f)

If you made a rule that blocks had to have transactions well people could just make a dummy transaction sending 1 satoshi to themselves.  

As the block subsidy declines and transaction fees become more important the "problem" will correct itself.  No need for the protocol to enforce it.

While it may not be the case today, in the past it was not uncommon for there to be periods of no transactions on the network [look at the first 30k blocks].  If the network refused to process blocks without a transaction, people would be forced to just insert dummy transactions in order to keep the block rate moving in proportion to the hash power of the network.  Otherwise, the difficulty could plummet as a result of low network activity, rather than low hash power.

Even today it is the case that 0 transaction blocks can come about due to -extremely- close blocks on the network.  However, eliminating these short blocks will end up offsetting the average block rate.  A 0-5 second block on the network smooths out the average when we have the occasional 1-hour without a block period.

So it's really two reasons that I see:

1) If 0-tx blocks were stopped, people would just send append a 0-fee transaction to their own blocks, which just wastes space in the chain.
2) If 0-tx blocks were stopped, the network difficulty would be slanted towards a higher difficulty than necessary [even if its not a major impact].

Thanks for the clarification.

In addition to the previous points every block empty or not makes it one block more work to rewrite a particular previous block making the tx in them a little bit more secure.

Very good point.  So anyone who's transactions were in the prior blocks benefits from increased depth in the blockchain.

It is based on time and amount (not just total amount but amounts of the prior output).

priority = sum(input_value_in_base_units * input_age)/size_in_bytes

You need a priority of 57,600 or higher to avoid spam fees.

IIRC (and I may not be) the simple version is 1 coin 1 day.

IF you receive 1 BTC you need to wait 1 day to avoid a fee when transferring it.
If you receive 10 BTC you need to wait 2.4 hours to avoid a fee when transferring it.

The wallet always attempts to reduce the fee using older coins as necessary to minimize or eliminate any fee.

it also depends on how your wallet is constituted... if all those 'old' transactions are millions of 0.00000001 inputs, then you'll still get fees because of the data size (in bytes) of your transaction.


edit: also, even if you had a wallet with 1000 BTC which was just ONE really old transaction... as soon as you spend 1 btc from it, the remaining 999 btc will be sent as 'change' and will no longer be an old transaction, but a really new one.

you can spend 0.00000001 without a fee if you have an old incoming 0.00000001 that was sent to your wallet.

No because once you send it, it is gone and the wallet where it is received it is now time 0.

best guess: yes (for ONE hour)

then you'd have the remaining 90% as 60 x 0.9 btc in new change addresses.

after the first hour (the first 60 x 0.1 transactions), all your wallet would contain would be new addresses with 0.9 in them.

No because you can't send part of a coin.

If you have a 1 BTC input and want to send 0.1 you can't do that 10 times.

You have to send 1 BTC so you send 0.1 and get 0.9 back as change.  The change now has a time 0 and is too young to be sent without incurring a fee.

sure, if they had RPC switched on i guess.

How many coins per transaction?

24 BTC?  You can send 24 BTC again in 6 confirmations without fee, since 6 confirmation is the minimum it could continue forever.  The point of the fees is to avoid cheaply spamming the network.  i.e. taking 1 BTC and sending a million 1 satoshi transactions and keep doing that every seconds until you generate TB worth of transactions.

The number of nodes is irrelivent.  It doesn't matter where you send coins.  Hell you can send them to another address in the same wallet.

Everytime you transfer a coin its time resets to 0.  If it is too young to be used in a new transaction you must pay a fee or wait.

Just to add to the talk about fees (in the case of 0tx blocks):  Don't forget that you can send your coins anywhere without a fee when you're the one mining a block :).  Or you could send .00000001 and tack on a 500 BTC fee, since you're the one mining that block [don't broadcast it, just include it in the block when you build it].

I don't get the fascination of 20 computers.  I already explained you could send it from 1 address in a wallet to another address in the SAME wallet and it would generate the same amount of transactions, same amount of transaction data, and incur the same limits.

Using 20 computers and 40 wallets would just be stupid.  Lots of electricity to accomplish nothing.


As far as transferring 24 BTC continually. It would create negligible transaction volume.  You must wait 6 confirms before spending so you could perform 1 transaction per hour, or ~24 per day.  At 0.25 KB per day you would add a massive 6KB per day or 2MB per year.

that's ridiculous in 2012... even an 'old' hard drive is 80gb+
a 'not old' hard drive is 500gb+.

Not without prohibitively massive cost.  That is the entire point of spam fee.  It prevents spam.   Well technically it doesn't prevent it.  If someone wanted to spend tens of millions of dollars acquiring bitcoins just to massively spam the the chain they could they would just lots tens of millions of dollars in wealth to transaction fees. :)  It would be a good day to be a miner.

Blockchain pruning could be used to clean up any "circular" spam.  To have a lasting effect you would need to send coins somewhere permanently (so it could be pruned out).  Thus costing you not just the spam fees but the principal.

TL/DR I wouldn't say impossible just prohibitively expensive.  You could also mine your own transactions "for free" but then you are limited to block size and how many blocks you can generate per day.  To generate a large number of blocks (full of spam) each day would require hundreds if not thousands of MH/s of hashing power (which isn't free). 

The later contradicts the former.

Couldn't be used for business except the fees would be lower.  How would lower fees make it impossible to be used by business.

Also you keep getting hung up on # of wallets.  # of wallets would have no effect on fees.  Amazon could have 1 wallet or a quadrillion wallets running on a quadrillion computers.  It would have absolutely no effect on fees, spam, transaction volume, etc.

actually d&t i think it would have a negative effect.

imagine a wallet has 3 transactions:

3 btc: old
3 btc: old
4 btc: new


you could make 2 x 1btc spends from that without incurring fees.

but let's say you had them in 3 different wallets:

a = 3 btc: old
b = 3 btc: old
c = 4 btc: new

if you spend from the 'c' wallet, you'd have to pay a fee because it can't find any old transactions to spend from... even though you've got plenty of old transactions in your other wallets.

Bitcoin isn't free.  It was never intended to be and never will be.  It is the largest distributed computing network in the world requiring hundreds of megawatts of power and millions of dollars in computing hardware.  What do you think pays for it?

Are you sure about that?  I thought it didn't.

Read CWallet::SelectCoinsMinConf() in wallet.cpp.  It tries to spend as little as possible, ignoring coin age.