Serial numbers

[acolombo]

No, they can tell exactly where they came from.

I read somewhere, I still can´t find it again, let´s suppose these transactions are performed, in this order:

 A --->  C  (1BTC)
 B --->  C  (1BTC)
 C --->  D  (1BTC)

Can you tell if D has received the coin once owned by A or by B?

[deepceleron]

I read somewhere, I still can´t find it again, let´s suppose these transactions are performed, in this order:

 A --->  C  (1BTC)
 B --->  C  (1BTC)
 C --->  D  (1BTC)

Can you tell if D has received the coin once owned by A or by B?

I would refer you to this post.

Like writing a check against my bank account, you can no longer say that the dollars in the check are the same dollars from my dwolla deposit or my direct-deposit paycheck. They don't have serial numbers any more in the bank's computer.

The only way you could make the statement that D received some of the currency funded by A would be if D > B + old balance, then the funds from A would be required to fulfill the transfer. Otherwise, there is a kind of deniability. You could also say that the input payments are spread out and co-mingled, so D is 1/2 of A and 1/2 of B, so it's all dirty money, etc.

The connections between different addresses are always transparent though, you can see which address has been sending to who.

[netrin]

Yeah, bitcoins really don't exist.

If I blow a kiss, does it exist? Even if I don't know, can't see, don't interact with the recipient? If the only record of the kiss is verified by people I've never met, who've never seen, touched, or in any way interacted with me directly, does the record prove the existence of said blown kiss to an anonymous potentially non-extant recipient? Existence is purely up to your own interpretation.

[kjj]

Pretty much all of you are confusing transactions, outputs and addresses.

There are no bitcoins, there are no addresses, there are no balances.  There are only transaction outputs.

In acolombo's scenario, yes, D can be traced back to either A or B, but not both (assuming no fee).

In a more complicated scenario, like if C wanted to send 1.5 BTC to D, he would have to redeem both A and B, and would get change back.  D would have coins that had come through both A and B, but we couldn't say that half were from A and half from B, or that 1 BTC was from A, and 0.5 was from B, or any other combination like that.  But we certainly can say that both A and B were both involved.

[acolombo]

Pretty much all of you are confusing transactions, outputs and addresses.

There are no bitcoins, there are no addresses, there are no balances.  There are only transaction outputs.

In acolombo's scenario, yes, D can be traced back to either A or B, but not both (assuming no fee).

In a more complicated scenario, like if C wanted to send 1.5 BTC to D, he would have to redeem both A and B, and would get change back.  D would have coins that had come through both A and B, but we couldn't say that half were from A and half from B, or that 1 BTC was from A, and 0.5 was from B, or any other combination like that.  But we certainly can say that both A and B were both involved.

But transactions must be from one address to another, then addresses do exist. And each address has a balance, I believe...

[kjj]

Pretty much all of you are confusing transactions, outputs and addresses.

There are no bitcoins, there are no addresses, there are no balances.  There are only transaction outputs.

In acolombo's scenario, yes, D can be traced back to either A or B, but not both (assuming no fee).

In a more complicated scenario, like if C wanted to send 1.5 BTC to D, he would have to redeem both A and B, and would get change back.  D would have coins that had come through both A and B, but we couldn't say that half were from A and half from B, or that 1 BTC was from A, and 0.5 was from B, or any other combination like that.  But we certainly can say that both A and B were both involved.

But transactions must be from one address to another, then addresses do exist. And each address has a balance, I believe...

Except that transactions are not from one address to another.  Transactions are from transaction outputs to scripts.

I don't mean for this to be a metaphysical discussion, like whether or not the color blue exists.  It clearly exists, except that it doesn't; in the same way with addresses and balances.  What I'm trying to do is help you understand the bitcoin system by explaining what parts of it are really there, and which parts have been imagined to help us work with it.

We can count up all of the unredeemed transaction outputs in the block chain that correspond to private keys that we control, and we can call that a balance, but there is no concept of a balance in the bitcoin system.  And we send coins to addresses, but addresses don't really exist in the system either, we just use them to create scripts.  And bitcoins don't really exist, but the system refuses to accept transactions where the outputs sum to more than the inputs, and we call the basic unit in those calculations a bitcoin.

We rarely deal with the system directly, but use metaphors to make it resemble things we already understand.  But the metaphor isn't the reality, and if you apply the metaphor too literally, you end up with a very incorrect understanding of the system.  So, sometimes it is necessary to look under the metaphor and see the naked system as it really is.

[acolombo]

Except that transactions are not from one address to another.  Transactions are from transaction outputs to scripts.

I don't mean for this to be a metaphysical discussion, like whether or not the color blue exists.  It clearly exists, except that it doesn't; in the same way with addresses and balances.  What I'm trying to do is help you understand the bitcoin system by explaining what parts of it are really there, and which parts have been imagined to help us work with it.

We can count up all of the unredeemed transaction outputs in the block chain that correspond to private keys that we control, and we can call that a balance, but there is no concept of a balance in the bitcoin system.  And we send coins to addresses, but addresses don't really exist in the system either, we just use them to create scripts.  And bitcoins don't really exist, but the system refuses to accept transactions where the outputs sum to more than the inputs, and we call the basic unit in those calculations a bitcoin.

We rarely deal with the system directly, but use metaphors to make it resemble things we already understand.  But the metaphor isn't the reality, and if you apply the metaphor too literally, you end up with a very incorrect understanding of the system.  So, sometimes it is necessary to look under the metaphor and see the naked system as it really is.

I think I get what you say, still addresses have a 1:1 relation to private keys, don´t they? OK, let´s say coins do not exist, but private keys do, then I beleive the client translates address to key and viceversa, got it right?

[netrin]

Kjj, a bitcoin address exists in so far as an email address or street address exists. What exists at these addresses are different levels of abstraction: a paper letter, a copy of text in a database, a signed record in a public balance sheet.

A bitcoin is only the identity one. One what? One bitcoin. One unit of value in a balanced public record. It has value because a community has agreed that an instance of this one bitcoin is unique. Its uniqueness is verified by the balancing of the public record (and faith that Euclid's 2300 year old assertions are true without proof).

A bitcoin's 'existence' is bound to a single spend solution. If a bitcoin could be copied then it would only be one of numerous instances of the number one. Only because the community forbids copying does a bitcoin 'exist'. In that way, a bitcoin is intellectual property. We may all know about it but not own it. We transfer ownership of mundane numbers by way of signed irreversible contracts/assertions.

One aspect I find very fascinating about bitcoins are that when we look at a typical balance sheet, the numbers represent other assets, dollars, widgets, etc. In the case of bitcoins, the numbers in the balance sheet are the bitcoins. If they represented other assets, those assets would be nowhere to be found. The record is what exists. The signed block chain ensures that it can't easily be duplicated nor manipulated.

Oh, and this should be a metaphysical discussion, because the original question (bitcoin serial numbers) has been thoroughly answered. Now we get to smoke cigars, drink cognac and discuss technological philosophy.

I think I get what you say, still addresses have a 1:1 relation to private keys, don´t they? (in all practicality, yes) OK, let´s say coins do not exist, but private keys do, then I beleive the client translates address to key and viceversa, got it right? (the address (hash) is derived from the public key which is derived from the private elliptic key, but not vice versa or in any other direction)

[Revalin]

Your BitCoins are just as real as your bank account balance: they're a few lines in a computer's database.  Neither has a concrete location: both BitCoin and your bank replicate the data widely, and withdraws may be performed anywhere where you present the proper credentials.

[netrin]

Your BitCoins are just as real as your bank account balance: they're a few lines in a computer's database.  Neither has a concrete location: both BitCoin and your bank replicate the data widely, and withdraws may be performed anywhere where you present the proper credentials.

Good point. And similarly I think we should drop the 'wallet' metaphor and call it a portfolio of accounts.

One difference though is that your bank account balance represents and can be redeemed for an equal quantity of a physical asset, dollars, euros, pesos, even if they don't actually exist 1:1 (1:10 in most countries). The bitcoin credentials are (nearly) the asset itself.

It's ironic though that bitcoins are not promissory notes, are redeemable for nothing, but yet neither are bitcoins fractionally diluted at all.  They truly are a new asset class.

[acolombo]

I think I get what you say, still addresses have a 1:1 relation to private keys, don´t they? (in all practicality, yes) OK, let´s say coins do not exist, but private keys do, then I beleive the client translates address to key and viceversa, got it right? (the address (hash) is derived from the public key which is derived from the private elliptic key, but not vice versa or in any other direction)

Of course the private key is not derived from the public data, but now again address relates 1:1 to public key. It´s interesting this thing, I still have my balance at 0.00, I suppose when I got my millions I will know better.

[netrin]

Of course the private key is not derived from the public data, but now again address relates 1:1 to public key. It´s interesting this thing, I still have my balance at 0.00, I suppose when I got my millions I will know better.

Yeah. Bitcoin is a wonderful system. I hope Satoshi will come out of hiding and accept his future Nobel Peace Prize!

Now, we can just say you've got it. And you do 99%... but just to be pedantic, for every address validly generated by the client, there will have existed a private key, which the human should not lose if he wishes to send bitcoins from that address's balance. I've not studied elliptic keys in detail, but if they are anything like RSA, each key of a pair uniquely satisfies Kpub(Kpriv(msg)) = msg = Kpriv(Kpub(msg)). However, there may be theoretically numerous messages which could produce the same SHA address, but we'd be extraordinarily unlikely to ever stumble upon such a collision before the oceans boil.

[acolombo]

Yeah. Bitcoin is a wonderful system. I hope Satoshi will come out of hiding and accept his future Nobel Peace Prize!

Now, we can just say you've got it. And you do 99%... but just to be pedantic, for every address validly generated by the client, there will have existed a private key, which the human should not lose if he wishes to send bitcoins from that address's balance. I've not studied elliptic keys in detail, but if they are anything like RSA, each key of a pair uniquely satisfies Kpub(Kpriv(msg)) = msg = Kpriv(Kpub(msg)). However, there may be theoretically numerous messages which could produce the same SHA address, but we'd be extraordinarily unlikely to ever stumble upon such a collision before the oceans boil.

Thanks a lot.

[Immunooel Gu]

Yes und nu.

Oovnersheep ooff a Beetcuin duesn't hefe-a tu be-a recurded. Bork bork bork! Su yuoo cuoold trece-a a Beetcuin boot it duesn't hefe-a tu hefe-a a teee-a tu un identeety, deffeeteeng uny use-a treceebeelity meeght hefe-a.