True.  Sending *from* the address is good too.

Also, the method doesn't have to be secure, since anyway it is not really very usefull to prove to someone taht you own some money.

Alice could prove Bob she owns enough money, but at the end the moment of truth occurs during the real transaction from Alice to Bob.  It's easy to prove you own any amount of money.  Bacically you could even borrow it during a short period of time.  Big money owners could also sell this kind of proof to people (selling the "proof", not the money).

What I mean is that proving you own some money is actualy quite useless in commercial relations anyway.  But if people really want to do it, they can do so with a relative security within the current bitoin implemenation.

True.  We actualy don't need it.

My initial post was intended to allow people to prove they own a private key.  It was intended to claim ownership of future paiements into the bitcoin address.  The idea was to ease a process a asset transfer.  It's completely different than proving you own a certain amount of money, which is in my opinion quite impossible  (since money could have been borrowed).

Also, proving you own some money doesn't prevent you from lying about your intent to spend it.

This functionality is redundant with that offered by the (existing, well understood, established, portable, and widely implemented) OpenPGP standard, and others. Saying that it isn't redundant because bitcoin doesn't already offer it doesn't make much sense. This kind of thinking is how the world ended up with atrocities like MS Outlook. I've said before that I like the Unix Way (TM) of doing things, and this is because there is just less room for disaster.

I bolded a section in your previous reply. My question stands. How can you ever establish an identity (such as when saying that you control a bitcoin account) without some external  bootstrapping mechanism? You can't ever "prove without a doubt" that you control a private key. The best you can do is convince someone to trust you to identify yourself correctly and not divulge your private key. I wonder if you understand the distinction that I am trying to draw.

And by bitcoin's nature, which you seem to think I never bothered to try to understand, you cannot keep secret the balance of an account after disclosing the ID. For an arbitrary account number (for which there is a record after a transaction), anyone can check its balance at any time, whether or not anyone knows who has ever controlled it.

Give it up, davout.  Gene is right.   It's a feature that is not necessary to bitcoin, and that could be implemented with an external, optionnal, program.

Therefore, there is no reason to add it in the official client.  It would give unecessary additional work for programmers and would add potential points of failure.

But I wish someone will eventually implement it as an external tool.

I guess.  However I keep thinking it should not be integrated inside the client but as some kind of a bitcoin-tools package, amongst which we could also find the explorer functions of blockexplorer.com.

I didn't know about gavin's tools.  I'll give it a look.

Why? I don't have to prove to a shopkeeper that I have enough money to just look around in his store. If I want to buy the item and I don't have enough money in my pockets, the transaction will fail at the point of sale. He loses nothing, and I lose nothing. This analogy generalizes, too. Again: credit vs. cash. Bitcoin is digital cash.

I have to agree with davout that exporting the private key is something useful in a scenario where you want to pay in bitcoins without having a bitcoin client with you (on a smart phone or other device with Internet access), only something like a smart card which holds securely the private key of one of your bitcoin addresses: it can sign the payment transaction. See this topic I opened: http://bt.irlbtc.com/view/2898.0

Gene: how are you able to pay some bitcoins with your PGP key at the store ? Without the private key of the bitcoin address, nobody can sign the transaction to validate it ! Or am I the one missing something here ?

But I can agree with grondilu that this can be part of external tools that manipulate your wallet.dat (on your PC) and make the transaction (at the merchant's store).

Going to take a look at gavin's tools if I have some spare time...

Here's where grondilu describes his scripts to sign arbitrary data with wallet keys using openssl. Very impressive.

http://bt.irlbtc.com/view/2694.msg39658#msg39658