Ten Minute Delay vs POS

[BitCoinDream]

When bitcoin is popular enough, there will be a few payment systems that will handle transactions off the blockchain. So, the average 10 minute block time will be irrelevant.

Won't we have the same problem that we have now with Paypal though; overpriced transaction fees because of too much middlemen?

Most of it would probably be automated. And there would be multiple companies competing against each other in the same space for lower fees. Have you ever had a Cryptsy account? You can use trade keys there to instantly transfer coins to other members with no fees and since the whole process is completely internal, all transfers are completely off the blockchain.

This would involve trusting a third party (cryptsy in your example) which is something that is generally against the principles of bitcoin. I think there is very little risk for merchants of a transaction not confirming (and/or a double spend attack) as long as the transaction is under a certain amount. A merchant could also have a rule that a TX must have at least "x" TX fee in order for them to accept it as 0/unconfirmed.

This does not solve the problem. Services like BitUndo can reverse Tx within 10 minutes even if Tx fee is paid.

[DoraTheBTCexplorer]

We need something, transactions are too slow 

[btchris]

This does not solve the problem. Services like BitUndo can reverse Tx within 10 minutes even if Tx fee is paid.

Services like BitUndo can attempt to reverse a tx before the next confirmed block, but they remain very unlikely to succeed (even as per their own FAQ). BitUndo depends on miners running a customized version of bitcoind, and thus far the interest in doing so has been pretty low.

[Pony789]

This does not solve the problem. Services like BitUndo can reverse Tx within 10 minutes even if Tx fee is paid.

Services like BitUndo can attempt to reverse a tx before the next confirmed block, but they remain very unlikely to succeed (even as per their own FAQ). BitUndo depends on miners running a customized version of bitcoind, and thus far the interest in doing so has been pretty low.

Basically the pool is just trying to find the next block with the "new tx" before any other pools finding the next block with the "original tx".
So unless that pool has a significant hashrate (say, 5% of total network hashrate), it is not really a big deal.

[tspacepilot]

Yes, I did mean point of sale. 

It looks to me that it would be possible to double spend/over spend if the goods/services are provided before the payment has cleared the blockchain.  For example, I buy a laptop at the Apple Store, I submit a payment (knowing my wallet doesn't have enough in it).  They see the pending transaction and I leave the store with my product.  Ten minutes later they find the transaction fails on the blockchain.

Right, if you were buying an expensive product I think they might want to see that you get a confirmation before they let you leave the store.  However, if you think about it, almost all restaurants let you consume the food before you pay anyway, no matter if its btc or cash you've already eaten by the time they give you the bill.  Just like you can send a bad transaction and then run, you can also just plain run without sending the bad transaction.  Restaurants already work on a "shame" system anyway where you just risk getting shamed if you don't pay.

[Mobius]

This does not solve the problem. Services like BitUndo can reverse Tx within 10 minutes even if Tx fee is paid.

Services like BitUndo can attempt to reverse a tx before the next confirmed block, but they remain very unlikely to succeed (even as per their own FAQ). BitUndo depends on miners running a customized version of bitcoind, and thus far the interest in doing so has been pretty low.

The reason their success rate has been so low is because their hashrate is low. As a result of their low hashrate they are able to mine few blocks. They are only able to reverse unconfirmed transactions when they mine a block prior to the TX being confirmed.