Hacker News new | ask | show | jobs
by tptacek 5501 days ago
Doesn't Bitcoin imply nonderministic and potentially large delays in transaction settlement?
2 comments

I didn't downvote you, but I'd answer yes and point out that the same criticism applies to email as a communications channel. And while you might not buy a house with a single BTC transaction, it's not that risky as a channel to send a modest payment to someone with a trustworthy reputation or with whom you have an ongoing relationship.

There are a lot of businesses that have to deal with outlandishly high rates charged by companies that process credit card payments for non-US companies. I'd find it far more convenient to send a bitcoin payment to a supplier in China than have to go to the bank and handle the conversion that way and would also love it if I could exchange BTC for HKD or CLP the next time I'm travelling in Hong Kong or Latin America.

Respectfully, that criticism applies to carrier pigeons too, which is why we don't use carrier pigeons or email for online transactions; we use payment processors that clear transactions in low single digit seconds.
The same criticism applies to cheques as well as COD, but I don't see anyone saying they don't serve a purpose.

Assuming BTC becomes a generic unit of exchange, the currency promises to save me around $50 to $100 per international money wire and cut the time it takes to transfer funds down to a few minutes. I've had zero failures using the system to date (max transfer time of 20 min), and would need to see a 10% failure rate before it stopped making sense. And that's without even calculating the costs of relying on monopolistic intermediaries like banks for currency exchange.

In a very technical sense, yes. And miners can choose to not include your transaction in the block (ie, if you don't meet the fee requirement they chose). It's entirely possible for your transaction to be continually pending and never accepted.

In practice, it takes within about 10 minutes for the first confirmation to go through, and 10 more for each one after that. Most merchants will likely be interested in waiting for a few more blocks to be added before totally accepting, but that can be solved by middle-men that provide insurance against double-spending, allowing immediate transactions. And transactions move across the network quite quickly - I tend to see them appear well within a minute.

I don't see why waiting for blocks is necessary. If a large portion of the network consenses within some small number of seconds, the likelihood that a rogue portion of the network would confirm conflicting transactions and then have that confirmation confirmed by that consensing network is vanishingly small.
In general, yes. But I do somewhat doubt that it'll be seen that way by many merchants - greater confidence they got paid is... greater confidence they got paid. Especially for internet purchases, a 10-20 minute delay is often unimportant.
Am I reading that right? On Internet purchases, 10 minute delays are unimportant? I chose who to buy Cisco routers from because of the specter of a 5-minute delay. Can you flesh that statement out a bit?
Essentially that the shipping time utterly dwarfs the validation delay. The main place this wouldn't be true is if you purchase an online activation key of some kind - if you purchased a downloadable, it'll often still take more than 10 minutes to complete.

The moment you send out your transaction (assuming it's valid), you can consider it approved, and move on with your life. The recipient can also acknowledge the transaction within a minute or so at most, it's the network validation and protection against alternate transaction histories that takes time.

FWIW, there's a pretty extensive discussion of the fast transaction problem here: http://forum.bitcoin.org/?topic=3441.0