Bitcoins

[BabylonHoruv]

http://www.bitlist.tk/

Looks like there are a few things up for sale now, mostly computer hardware.

[Requia ☣]

Hmm, I may have screwed this up on my price.

it seems the other things being sold there suggest a BTC value far higher than what the exchange services are paying.

[Cain]

homepage is working fine for me.  Maybe it is being blocked from Europe?

Doubt it, GA is Amurrican based, IIRC.  Also, what Requia said.

[BabylonHoruv]

Hmm, I may have screwed this up on my price.

it seems the other things being sold there suggest a BTC value far higher than what the exchange services are paying.

I think it is more likely the others screwed up.  bitcoinmarket.com is a pretty good indicator of the dollar value of bitcoins.

[Mesozoic Mister Nigel]

I keep reading the title of this as "bitchcoins".

[BabylonHoruv]

heh,  might have some trouble trading in those.  On the other hand they might get awfully popular.

[Golden Applesauce]

I'm a little more optimistic about the idea now, after reading that PDF.  (The site works again now.)  I still have some concerns, though.  Treat everything below as moderately uninformed speculation; my only credentials are knowing what the words "hash," "public/private key," and "computational complexity" mean, and I still don't fully understand their system.

The main security comes from the assumption that the vast majority of nodes hashing the transaction history are honest.  Probably okay, but the hidden assumption is that the network is homogeneous.  If I try to buy a widget from a seller in the Russian network and from a seller in the US network with the same coins, two things can happen.  If the Russian network and the US network are heavily interconnected, then one transaction will make it into a block first and both networks will count the larger chain as legit.  If they are only loosely interconnected, then the transaction sent to the Russian network will get added to a chain that's kind of local to the Russian chain and the US transaction will get added to a chain kind of local to the US.  Eventually the Russian chain and the US chain will get compared and the shorter chain will get cut - invalidating one of my transactions, but possibly after the seller had already mailed me the widget.  Would that also invalidate every transaction that was only on the fork of the chain that got shorter?  I need to research their system more.

It could also be possible to 'spoof' to a target node and masquerade as an entire network.  Suppose Alice buys a widget from Bob; Bob waits to see that enough other nodes accept the transaction before acting on it by mailing me the widget.  But Alice has somehow cut Bob out of the main network; Bob thinks he's talking to a random sample of nodes, but in fact a huge proportion of them are Alice's sockpuppets.  Is that viable?

If the network is too anonymous, I could see other problems.  If I understand correctly, the only thing that happens if you try (and fail) to double spend your coins is that one of the transactions falls through - which is no worse for you than if you hadn't tried to cheat in the first place.  So even if there's only a small chance of getting away with double spending, as long as trying doesn't cost you anything, a huge proportion of people are going to try it.  There must be some way of providing for a mechanism of identifying known cheaters - but to do that you'd need to be able to link multiple transactions to the same account (once could be an accident, more attempts would be intentional.)  But if you can link multiple transactions to the same account, then you can deduce the transaction history of people and start putting together PI.  Not sure how you can detect cheaters without being able to figure out any given person's transaction history.  Also, since failed attempts at cheating are simply discarded, is anyone (besides the victim) even going to know if there was an attempt?  Keeping track of failed transactions doesn't help you hash new transactions, so it isn't profitable - someone would need to be payed to do it, probably.

The main thing, though, is that people would be using this system on their home computers.  If I understand correctly, if someone else gets a hold of your private key, you're toast and there's nothing you can do about it.  How well is the typical user going to be able to protect his data?  There's a reason banks pay the big bucks for computer security professionals - it's necessary when the stakes are people's life savings.  I think my computer is reasonably secure, but I wouldn't bet my college tuition on it.