5 ms·

There was a visible fork of the blockchain - a clearly exceptional condition showing that the network did not work correctly at that moment. All merchants shou

by zby 14y ago

There was a visible fork of the blockchain - a clearly exceptional condition showing that the network did not work correctly at that moment. All merchants should stop accepting payments for duration of that - this can easily be automated.

URSpider94 14y ago
Forks of the blockchain happen all the time as a natural side effect of the distributed nature of blockchain calculation. How do you propose that a merchant detect a pathological fork? If you're going to tell me that it depends upon the length of the fork, then you'd better be waiting at least that many blocks before you verify a transaction.
- nazgulnarsil 14y ago
  this is already built into the protocol.
  - URSpider94 14y ago
    what is 'this'? If you mean waiting for x validations before accepting a transaction, yes -- but do you really think that it's practical to wait 30 minutes before delivering goods for a transaction? That's going to be a really long time spent staring at the little spinny thing over a line of text saying "verifying your payment" ...
    - nazgulnarsil 14y ago
      It's completely reasonable. Bitcoin is not a good candidate for a medium of exchange despite what some adopters will tell you. It is more like virtual gold. You can see in the real world that mediums of exchange and stores of value aren't the same thing because the properties that are desirable in one are not desirable in the other.
      - eric_bullington 14y ago
        And there's another electronic currency system coming online now that should handle instant transactions well: ripple.com Its a new incarnation of a relatively old idea, but the good thing about the new ripple is that transactions are (and should continue to be) nearly instant. And it is very, very easy to trade bitcoin for ripple using the ripple system, and should get even easier. So I imagine a future where value is stored in btc, and spent using ripple (in any of the supported currencies).
    - johnsoft 14y ago
      There are services out there that will absorb transaction risks for you. One example is https://bitpay.com/ https://bitpay.com/
    - runn1ng 14y ago
      Even if they did wait - if the pathological chain was longer, than 6 blocks, it wouldn't help them much.
    - lotsofcows 14y ago
      How do you think credit card transactions work? The auth code you get back from a "valid" cc transaction doesn't represent a transfer of money. That might not happen for another 24 hours.
      - Nursie 14y ago
        Sure, but the 'authorisation' portion of a CC transaction reserves the funds and more or less guarantees payment.
        lotsofcows 14y ago
        Doesn't reserve funds, it's rarely realtime. "Guarantees payment" just means that the cost has been shifted from the retailer to the card provider - for a while anyway, then the card provider starts grabbing money back.
        Nursie 14y ago
        Yeah it does reserve the funds, I've worked in this industry. Settlement happens later, but authorisation does put a reserve on the funds in the credit account if the transaction has gone end-to-end. If you mean the 'chargeback' feature when you say "then the card provider starts grabbing money back", then yes, absolutely, this is a feature of card payments and is essential for people to have any trust in internet commerce at all. Chargeback is good
        lotsofcows 14y ago
        End to end is the key point, loads of places still use off line terminals that batch the job up at eod. I wasn't referring to chargeback; I meant stroppy card providers. The sort of thing these guy are fighting (although that was for a security breach rather than too many authorised charges against empty accounts): http://www.out-law.com/en/articles/2013/march/retail-victim-of-cyber-attack-in-bid-to-recover-penalties-levied-by-visa-over-alleged-pci-dss-rules-breach/ http://www.out-law.com/en/articles/2013/march/retail-victim-...
- zby 14y ago
  A pathological fork is when both parts grow in a visible way. The normal thing is when a part of the network is split off - and it grows it's own blockchain - then it gets connected and the chains get re-conciliated. When the reconciliation does not happen even if it could - that is when it is visible that both chains grow - then either the internet is in some funny state or the bitcoin system is.
rmc 14y ago
this can easily be automated. No it can't. The only way is to have a source saying "Bitcoin is OK, accept transactions" or "Bitcoin is not OK, no transactions". However this source now will be the target for no end of attacks (both hacking and legal), since it literally allows you to turn off the bitcoin economy. The bitcoin industry/community seems incapable of writing secure software. So I wouldn't put much faith in them doing anything well.
- RoboTeddy 14y ago
  > However this source now will be the target for no end of attacks (both hacking and legal) It's possible to build a decentralized fork detector that anyone can run. All you have to do is run a bunch of versions of bitcoin nodes, and run a cron job to check and see if they have differing opinions on which fork is currently valid. If you discover a recent fork of length > x, you might want to hold off on verifying or making transactions for a bit until things clear up. > The bitcoin industry/community seems incapable of writing secure software Dan Kaminsky (http://en.wikipedia.org/wiki/Dan_Kaminsky http://en.wikipedia.org/wiki/Dan_Kaminsky) would probably disagree: "When I first looked at the code, I was sure I was going to be able to break it, Kaminsky said, noting that the programming style was dense and inscrutable. The way the whole thing was formatted was insane. Only the most paranoid, painstaking coder in the world could avoid making mistakes.…He quickly identified nine ways to compromise the system…when he found the right spot, there was a message waiting for him. Attack Removed, it said. The same thing happened over and over, infuriating Kaminsky. I came up with beautiful bugs, he said. But every time I went after the code there was a line that addressed the problem.…I’ve never seen anything like it, Kaminsky said, still in awe…Either there’s a team of people who worked on this, Kaminsky said, or this guy is a genius." Kaminsky was talking about Satoshi's code, but some of the current devs seem good. For example, there's Mike Hearn, a senior security engineer at Google. And Jeff Garzik, a linux kernel developer.
  - Nursie 14y ago
    The bitcoin industry/community is not the same as the official client though is it? I'm pretty sure that the poster was referring to the community of services that exist around bitcoin, rather than the protocol itself, that have been repeatedly and expensively compromised.

by zby 14y ago

URSpider94 14y ago

Forks of the blockchain happen all the time as a natural side effect of the distributed nature of blockchain calculation. How do you propose that a merchant detect a pathological fork? If you're going to tell me that it depends upon the length of the fork, then you'd better be waiting at least that many blocks before you verify a transaction.

nazgulnarsil 14y ago
this is already built into the protocol.
- URSpider94 14y ago
  what is 'this'? If you mean waiting for x validations before accepting a transaction, yes -- but do you really think that it's practical to wait 30 minutes before delivering goods for a transaction? That's going to be a really long time spent staring at the little spinny thing over a line of text saying "verifying your payment" ...
  - nazgulnarsil 14y ago
    It's completely reasonable. Bitcoin is not a good candidate for a medium of exchange despite what some adopters will tell you. It is more like virtual gold. You can see in the real world that mediums of exchange and stores of value aren't the same thing because the properties that are desirable in one are not desirable in the other.
    - eric_bullington 14y ago
      And there's another electronic currency system coming online now that should handle instant transactions well: ripple.com Its a new incarnation of a relatively old idea, but the good thing about the new ripple is that transactions are (and should continue to be) nearly instant. And it is very, very easy to trade bitcoin for ripple using the ripple system, and should get even easier. So I imagine a future where value is stored in btc, and spent using ripple (in any of the supported currencies).
  - johnsoft 14y ago
    There are services out there that will absorb transaction risks for you. One example is https://bitpay.com/ https://bitpay.com/
  - runn1ng 14y ago
    Even if they did wait - if the pathological chain was longer, than 6 blocks, it wouldn't help them much.
  - lotsofcows 14y ago
    How do you think credit card transactions work? The auth code you get back from a "valid" cc transaction doesn't represent a transfer of money. That might not happen for another 24 hours.
    - Nursie 14y ago
      Sure, but the 'authorisation' portion of a CC transaction reserves the funds and more or less guarantees payment.
      - lotsofcows 14y ago
        Doesn't reserve funds, it's rarely realtime. "Guarantees payment" just means that the cost has been shifted from the retailer to the card provider - for a while anyway, then the card provider starts grabbing money back.
        Nursie 14y ago
        Yeah it does reserve the funds, I've worked in this industry. Settlement happens later, but authorisation does put a reserve on the funds in the credit account if the transaction has gone end-to-end. If you mean the 'chargeback' feature when you say "then the card provider starts grabbing money back", then yes, absolutely, this is a feature of card payments and is essential for people to have any trust in internet commerce at all. Chargeback is good
        lotsofcows 14y ago
        End to end is the key point, loads of places still use off line terminals that batch the job up at eod. I wasn't referring to chargeback; I meant stroppy card providers. The sort of thing these guy are fighting (although that was for a security breach rather than too many authorised charges against empty accounts): http://www.out-law.com/en/articles/2013/march/retail-victim-of-cyber-attack-in-bid-to-recover-penalties-levied-by-visa-over-alleged-pci-dss-rules-breach/ http://www.out-law.com/en/articles/2013/march/retail-victim-...
zby 14y ago
A pathological fork is when both parts grow in a visible way. The normal thing is when a part of the network is split off - and it grows it's own blockchain - then it gets connected and the chains get re-conciliated. When the reconciliation does not happen even if it could - that is when it is visible that both chains grow - then either the internet is in some funny state or the bitcoin system is.

rmc 14y ago

this can easily be automated. No it can't. The only way is to have a source saying "Bitcoin is OK, accept transactions" or "Bitcoin is not OK, no transactions". However this source now will be the target for no end of attacks (both hacking and legal), since it literally allows you to turn off the bitcoin economy. The bitcoin industry/community seems incapable of writing secure software. So I wouldn't put much faith in them doing anything well.

RoboTeddy 14y ago
> However this source now will be the target for no end of attacks (both hacking and legal) It's possible to build a decentralized fork detector that anyone can run. All you have to do is run a bunch of versions of bitcoin nodes, and run a cron job to check and see if they have differing opinions on which fork is currently valid. If you discover a recent fork of length > x, you might want to hold off on verifying or making transactions for a bit until things clear up. > The bitcoin industry/community seems incapable of writing secure software Dan Kaminsky (http://en.wikipedia.org/wiki/Dan_Kaminsky http://en.wikipedia.org/wiki/Dan_Kaminsky) would probably disagree: "When I first looked at the code, I was sure I was going to be able to break it, Kaminsky said, noting that the programming style was dense and inscrutable. The way the whole thing was formatted was insane. Only the most paranoid, painstaking coder in the world could avoid making mistakes.…He quickly identified nine ways to compromise the system…when he found the right spot, there was a message waiting for him. Attack Removed, it said. The same thing happened over and over, infuriating Kaminsky. I came up with beautiful bugs, he said. But every time I went after the code there was a line that addressed the problem.…I’ve never seen anything like it, Kaminsky said, still in awe…Either there’s a team of people who worked on this, Kaminsky said, or this guy is a genius." Kaminsky was talking about Satoshi's code, but some of the current devs seem good. For example, there's Mike Hearn, a senior security engineer at Google. And Jeff Garzik, a linux kernel developer.
- Nursie 14y ago
  The bitcoin industry/community is not the same as the official client though is it? I'm pretty sure that the poster was referring to the community of services that exist around bitcoin, rather than the protocol itself, that have been repeatedly and expensively compromised.