Blocks are still taking way too long. - page 2.

Este Nuno

legendary

Activity: 826

Merit: 1002

amarha

Quote from: Sheldor333 on October 18, 2014, 06:44:25 AM

I do think that could be improved if it was about 10 seconds each time that would help a lot in marketing it but then I think most people know that.

Ethereum is going to attempt 12 second confirmations from what I understand. https://blog.ethereum.org/2014/07/11/toward-a-12-second-block-time/

Will Bitcoin be able to adapt that technology with our current system or is this something that requires a full rewrite from the ground up?

Sheldor333

sr. member

Activity: 406

Merit: 250

I do think that could be improved if it was about 10 seconds each time that would help a lot in marketing it but then I think most people know that.

Dabs

legendary

Activity: 3416

Merit: 1912

The Concierge of Crypto

Quote from: the joint on October 17, 2014, 07:29:29 PM

With BTC, I can give you a few models which remove the need for checkout lines, checkout counters, and even cashiers, and only require a couple security personnel at the doors Cheesy

I'm listening. (or reading.) Go ahead, tell us about it.

the joint

legendary

Activity: 1834

Merit: 1020

Quote from: sakira on October 18, 2014, 03:03:02 AM

Quote from: Bit_Happy on October 11, 2014, 12:10:53 AM

Over 30 minutes for a single confirmation does not (and will not ever) work in the real world.
(Almost) Every time I start to believe that BTC is "going mainstream", I have a delay waiting for a 30+ minute block.... Shocked

still fortunately you only 30 minutes, I was almost 1 hour more because my connection is only 20KBps. moreover offline wallet so many blocks

That isn't how confirmations work. Confirmations happen through mining (e.g. 1 'solved block' = 1 confirmation). Additionally, 10 minutes, 30 minutes, or even an hour is just fine for most online transactions and a majority, if not most, non-micropayments (e.g. paychecks).

ranochigo

legendary

Activity: 3038

Merit: 4418

Crypto Swap Exchange

Quote from: sakira on October 18, 2014, 03:03:02 AM

Quote from: Bit_Happy on October 11, 2014, 12:10:53 AM

Over 30 minutes for a single confirmation does not (and will not ever) work in the real world.
(Almost) Every time I start to believe that BTC is "going mainstream", I have a delay waiting for a 30+ minute block.... Shocked

still fortunately you only 30 minutes, I was almost 1 hour more because my connection is only 20KBps. moreover offline wallet so many blocks

I don't think confirmation relies on connection speed. Transaction is broadcasted to the network almost instantaneously regardless of connection speed, of cause, slower connection speed may have delays but it isn't much. I think you are talking about syncing of blocks. It is also dependent on connection speed of your peers.

sakira

hero member

Activity: 532

Merit: 500

Quote from: Bit_Happy on October 11, 2014, 12:10:53 AM

Over 30 minutes for a single confirmation does not (and will not ever) work in the real world.
(Almost) Every time I start to believe that BTC is "going mainstream", I have a delay waiting for a 30+ minute block.... Shocked

still fortunately you only 30 minutes, I was almost 1 hour more because my connection is only 20KBps. moreover offline wallet so many blocks

Este Nuno

legendary

Activity: 826

Merit: 1002

amarha

Quote from: Dabs on October 17, 2014, 06:27:30 PM

Quote from: the joint on October 17, 2014, 02:05:11 PM

I think it's asinine to use the Starbucks example when a grocery store checkout line is a more Universal scenario. The products aren't paid for until after they are scanned and a final price is reached. Now you have all those people waiting in line...

There is a supermarket chain here (I actually shop there occasionally) that has it's own credit card. It's not Visa or Mastercard, it's their own card. Sort of like a membership card. You are assigned a credit limit, and they bill you for your purchases that month.

That's a real life example of how an establishment can use it's own version of off-chain transactions.

Either you pay the monthly or weekly bill in BTC, or you pre-pay and load up the card with BTC.

I could see this happening on a large scale in a future dominated by a handful of corporate groups. Like you have your Starbucks card, Walmart card, McDonalds card all bought with BTC at a discount. That is sort of happening already with Gyft. But it kind of breaks down in economies not dominated by a few mega corps. No one wants hundreds of cards. Perhaps that can be solved with small businesses groups creating their own cards or something.

the joint

legendary

Activity: 1834

Merit: 1020

Quote from: Dabs on October 17, 2014, 06:27:30 PM

Quote from: the joint on October 17, 2014, 02:05:11 PM

I think it's asinine to use the Starbucks example when a grocery store checkout line is a more Universal scenario. The products aren't paid for until after they are scanned and a final price is reached. Now you have all those people waiting in line...

There is a supermarket chain here (I actually shop there occasionally) that has it's own credit card. It's not Visa or Mastercard, it's their own card. Sort of like a membership card. You are assigned a credit limit, and they bill you for your purchases that month.

That's a real life example of how an establishment can use it's own version of off-chain transactions.

Either you pay the monthly or weekly bill in BTC, or you pre-pay and load up the card with BTC.

I think we can do better than that.

With BTC, I can give you a few models which remove the need for checkout lines, checkout counters, and even cashiers, and only require a couple security personnel at the doors Cheesy

Dabs

legendary

Activity: 3416

Merit: 1912

The Concierge of Crypto

Quote from: the joint on October 17, 2014, 02:05:11 PM

I think it's asinine to use the Starbucks example when a grocery store checkout line is a more Universal scenario. The products aren't paid for until after they are scanned and a final price is reached. Now you have all those people waiting in line...

There is a supermarket chain here (I actually shop there occasionally) that has it's own credit card. It's not Visa or Mastercard, it's their own card. Sort of like a membership card. You are assigned a credit limit, and they bill you for your purchases that month.

That's a real life example of how an establishment can use it's own version of off-chain transactions.

Either you pay the monthly or weekly bill in BTC, or you pre-pay and load up the card with BTC.

the joint

legendary

Activity: 1834

Merit: 1020

Quote from: Este Nuno on October 17, 2014, 09:08:09 AM

Quote from: Gemminyc on October 17, 2014, 08:53:52 AM

Those users who preffer faster confirmation will never be happy even with 30 second blocks, because sometimes you have to wait many minutes (even 10min) because of variance. And the orphaned blocks rate become really problem.

The point is to trust 0 confirm transactions, but some 0 confirm transactions are risky (no fee, small fee, using unconfirmed input), other 0 confirm transactions are much more secure - the one withhout unconfirmed input and above average fee

Yeah, the current Bitcoin based cryptocurrency system is definitely far from perfect, but the balance between a reasonable orphan rate and user friendliness probably lies somewhere between 1-2.5 mins. Given that 2 minute confirms have been thoroughly tested now on other platforms I think it's clear that it would be a positive change. The least viable confirmation time is definitely up for debate, but I don't really see any advantage to 10 minute confirms unless I'm looking at it from the perspective of a miner. And even then, if an upgrade like that can benefit users it might increase demand which could offset miners losses via orphaned blocks.

Quote from: Dabs on October 15, 2014, 09:43:15 PM

I found this from a much older thread.

Quote from: DeathAndTaxes on May 03, 2012, 10:31:49 AM

The 0-confirm double spend is something which has been known and studies for some time.

That being said a 10 sec window for real world face to face transactions is just asinine.

Counter experiment.
Go to starbucks, order, swipe your credit card.
START STOPWATCH
wait for coffee ... wait .... wait ... get coffee ... walk to door ... get in car ... drive out of line of site.
STOP STOP WATCH

I am 99% certain the elapsed time is going to be >10 sec.

There are methods to reduce (but likely never eliminate) the fraud risk of 0-confirm double spends. That being said no payment system is without risk.

Cash: counterfeit bills, theft, bank processing fees (which is why stores "generously" allow cashback)
Credit cards: stolen cards, friendly fraud, never ending cut to VISA/MC.

The goal isn't to reduce fraud/theft/loss to 0%. Doing that likely will drive away consumers. The goal is to reduce the COST OF PROCESSING transactions which includes everything from employee training to fraud to bank fees to infrastructure/logistics.

I think it's asinine to use the Starbucks example when a grocery store checkout line is a more Universal scenario. The products aren't paid for until after they are scanned and a final price is reached. Now you have all those people waiting in line...

And, just because the goal isn't to achieve a 0% risk currency due to near-impossibility doesn't mean that decreasing risk isn't important whatsoever, or that we shouldn't consider the indirect implications of slower transactions. For example, more time in the checkout line --> fewer customers overall --> less revenue generated --> smaller profit margin.

My biggest gripe with these blanket assumptions about BTC's technical specs is that people constantly forget the psychological implications. If you make a decision that even affects 2-3% of a company's consumer base, all of a sudden you're playing roulette with future revenue.

Este Nuno

legendary

Activity: 826

Merit: 1002

amarha

Quote from: Gemminyc on October 17, 2014, 08:53:52 AM

Those users who preffer faster confirmation will never be happy even with 30 second blocks, because sometimes you have to wait many minutes (even 10min) because of variance. And the orphaned blocks rate become really problem.

The point is to trust 0 confirm transactions, but some 0 confirm transactions are risky (no fee, small fee, using unconfirmed input), other 0 confirm transactions are much more secure - the one withhout unconfirmed input and above average fee

Yeah, the current Bitcoin based cryptocurrency system is definitely far from perfect, but the balance between a reasonable orphan rate and user friendliness probably lies somewhere between 1-2.5 mins. Given that 2 minute confirms have been thoroughly tested now on other platforms I think it's clear that it would be a positive change. The least viable confirmation time is definitely up for debate, but I don't really see any advantage to 10 minute confirms unless I'm looking at it from the perspective of a miner. And even then, if an upgrade like that can benefit users it might increase demand which could offset miners losses via orphaned blocks.

Gemminyc

full member

Activity: 158

Merit: 100

Those users who preffer faster confirmation will never be happy even with 30 second blocks, because sometimes you have to wait many minutes (even 10min) because of variance. And the orphaned blocks rate become really problem.

The point is to trust 0 confirm transactions, but some 0 confirm transactions are risky (no fee, small fee, using unconfirmed input), other 0 confirm transactions are much more secure - the one withhout unconfirmed input and above average fee

Este Nuno

legendary

Activity: 826

Merit: 1002

amarha

Quote from: ranochigo on October 17, 2014, 05:49:22 AM

Quote from: Pulley3 on October 17, 2014, 05:11:41 AM

Quote from: hua_hui on October 11, 2014, 03:05:34 AM

If you are not patient, you can go to other altcoins having quick confirmation, such as NXt, doge etc.

NXT and doge is still way way behind bitcoin. Not gonna do that.

I agree with this fully. We often see cryptocoins on news but usually, the one getting the media's attention is Bitcoin. Alt coin is usually more volatile in price than Bitcoin. Unless there is a service which immediately exchanges your coins to fiat or Bitcoin without losing the value if the market crash, I don't see why merchants who are using Bitpay would switch to alt coins. Bitpay already accepts instant payment without confirmation for most cases.

Bitpay could easily add payment options for an alt that was backwards compatible with Bitcoin. That is of course if it made sense for them financially and the alt was liquid enough.

ranochigo

legendary

Activity: 3038

Merit: 4418

Crypto Swap Exchange

Quote from: Pulley3 on October 17, 2014, 05:11:41 AM

Quote from: hua_hui on October 11, 2014, 03:05:34 AM

If you are not patient, you can go to other altcoins having quick confirmation, such as NXt, doge etc.

NXT and doge is still way way behind bitcoin. Not gonna do that.

I agree with this fully. We often see cryptocoins on news but usually, the one getting the media's attention is Bitcoin. Alt coin is usually more volatile in price than Bitcoin. Unless there is a service which immediately exchanges your coins to fiat or Bitcoin without losing the value if the market crash, I don't see why merchants who are using Bitpay would switch to alt coins. Bitpay already accepts instant payment without confirmation for most cases.

Pulley3

newbie

Activity: 25

Merit: 0

Quote from: hua_hui on October 11, 2014, 03:05:34 AM

If you are not patient, you can go to other altcoins having quick confirmation, such as NXt, doge etc.

NXT and doge is still way way behind bitcoin. Not gonna do that.

Este Nuno

legendary

Activity: 826

Merit: 1002

amarha

Quote from: Argwai96 on October 16, 2014, 06:15:15 PM

Quote from: Este Nuno on October 15, 2014, 02:41:11 AM

Quote from: OrientA on October 14, 2014, 03:42:28 AM

Quote from: Este Nuno on October 13, 2014, 06:16:31 AM

So I wonder why people haven't been asking the core team to change the confirm time on the next hard fork? This is a perfect time to do it considering we're very likely going to be doing a hard fork any way to increase the block size from 1 mb.

It takes about 1 hour to make sure the confirmation is solid due to orphan blocks or some kind of attacks. 6 confirmation of 10 minute block takes 1 hour. If you reduce the block time to 1 min, then it will take 60 confirmation to avoid double spend.

Right, so we might as well have shorter confirms since no one is going to be rewriting blocks on the Bitcoin chain for any transaction that's worth less than the cost to do so. And for the ones that are, we wait an hour. Sounds pretty good to me.

Quote from: FattyMcButterpants on October 14, 2014, 11:55:05 PM

Quote from: the joint on October 13, 2014, 03:54:34 AM

Quote from: Este Nuno on October 13, 2014, 03:41:31 AM

Quote from: mnmShadyBTC on October 12, 2014, 09:07:06 PM

Quote from: jonald_fyookball on October 12, 2014, 04:43:56 PM

Quote from: mnmShadyBTC on October 12, 2014, 04:41:37 PM

Quote from: leopard2 on October 12, 2014, 02:05:49 PM

Quote from: franky1 on October 11, 2014, 03:10:39 AM

Quote from: Bit_Happy on October 11, 2014, 12:10:53 AM

Over 30 minutes for a single confirmation does not (and will not ever) work in the real world.
(Almost) Every time I start to believe that BTC is "going mainstream", I have a delay waiting for a 30+ minute block.... Shocked

solution:
if your paying for a coffee or a mars bar, dont wait. if your buying a house wait 30 minutes

solution:
if your paying for a coffee or a mars bar, use LTC or DOGE. if your buying a house wait 30 minutes Wink

One confirmation with LTC or DOGE is much less secure then a 0/unconfirmed TX with bitcoin. Both of their network is not secure so it would be easy to attack the network and reverse your TX. Once a TX is propagated throughout the network, nodes will not accept TXs that use the same inputs so you will not be able to broadcast a TX that spends the same inputs, the only way to double spend is to confirm a block that uses at least one of the same inputs as what you just spent. As long as the value of your cup of coffee is less then what it costs to mine the next block then you are almost certainly safe to accept a 0/unconfirmed TX after the TX has been properly propigated

Not disagreeing with you , but isnt LTC fairly secure?

It is the most secure altcoin out there as of now but the short block time gives a misleading picture as to how secure it's network is. Even if it has the same difficulty as bitcoin it's 2.5 minute block time means that it takes 25% of the work to find one block and there will be at least 4x as many orphaned blocks.

If you want to do an apple to apples comparison of network security the LTC network has a difficulty of ~36.5 thousands now, compared to 35 million for bitcoin. Granted BTC and LTC miners are not interchangeable but I think this gives a pretty good comparison as to how easy it would be to attack either network.

The thing is that one confirmation is more secure than zero confirmations as it prevents a whole category of zero confirmation double spend attacks. So that confirmation comparison you're making is only accurate for a comparison of attacks where a rogue miner is actually publishing blocks. And the amount of time you need to wait for the same amount of security is the same so they're actually equivalent. Like I said up thread, the only disadvantage to shorter block times seems to be more orphans, but since it's something that benefits the end user I would think most people who aren't miners would actually be behind such a change. And the smart ones who are miners will see that despite the fact that they will be wasting some hash in some cases, the potential gains in demand and positive perception due to people not having to wait around 30 minutes for a confirmation often enough should make up for it in the long run. That's just conjecture of course, but considering the amount of complaints we hear about Bitcoin's slow block targets I wouldn't be surprised if that was the case.

It's not just more secure, it's infinitely more secure.

Fast-confirm coins absolutely have a niche, in my opinion.

Fast confirm coins are much less secure and having one confirmation on a fast confirm network is less secure then a 0/unconfirmed TX on a 10 minute network. The reason for this is that nodes will not accept a double spend transaction when there is a propagated TX that spends certain inputs. As long as nodes are sufficiently diverse and TXs are able to propagate quickly, then it will not be possible to double spend certain inputs you have spent on a 0/unconfirmed TX unless you find a block and include the double spend TX in that found block. At the very least you will need to wait for enough confirmations to equal one "10 minute" confirmations to get the same security of a 0/unconfirmed TX assuming the hashrate is the same.

The problem with this is that shorter confirmation blockchains also give miners more of an incentive to mine on shorter blockchains to prevent their chain from being orphaned. This will result in more orphaned chains and longer chains, ones that will potentially exceed the equivalent of a 10 minute confirmation

But it is possible to double spend unconfirmed transactions if you know how to do it correctly(not saying it's easy). Which is why the 1 confirmation is better, since that whole category of double spend attacks that are possible are protected against with one confirm.

I think at this point it's pretty clear that Bitcoin could upgrade to 2 minute blocks on the next hard fork and people would be happy with the change. One confirmation, even if it's 1/5th as secure as a one 10 minute confirm is still very secure on the Bitcoin network. I think it would beneficial for Bitcoin to make this change.

Here's Peter Todd's take on unconfirmed transactions: http://www.reddit.com/r/Bitcoin/comments/239bj1/doublespending_unconfirmed_transactions_is_a_lot/

Note: I'm not saying that unconfirmed transactions aren't fine for microtransactions and such.

Changing to 2 minute blocks would not be an upgrade. The only way to double spend an unconfirmed transaction if you do not control the miner that finds a block that confirms a double spend TX is to control a well connected node that miners connect to (you would probably actually need to control several well connected nodes)

How would it not be an upgrade to receive one confirmation that provides 1/5th of the security in 1/5th of the time? When 1/5th of the security is very significant and sufficient for all but the highest value transactions.

The cost benefit trade off seems pretty clear to me. A bit more wasted hashing shouldn't be a concern when considering increasing the usability of Bitcoin.

Argwai96

legendary

Activity: 1036

Merit: 1000

Thug for life!

Quote from: Este Nuno on October 15, 2014, 02:41:11 AM

Quote from: OrientA on October 14, 2014, 03:42:28 AM

Quote from: Este Nuno on October 13, 2014, 06:16:31 AM

So I wonder why people haven't been asking the core team to change the confirm time on the next hard fork? This is a perfect time to do it considering we're very likely going to be doing a hard fork any way to increase the block size from 1 mb.

It takes about 1 hour to make sure the confirmation is solid due to orphan blocks or some kind of attacks. 6 confirmation of 10 minute block takes 1 hour. If you reduce the block time to 1 min, then it will take 60 confirmation to avoid double spend.

Right, so we might as well have shorter confirms since no one is going to be rewriting blocks on the Bitcoin chain for any transaction that's worth less than the cost to do so. And for the ones that are, we wait an hour. Sounds pretty good to me.

Quote from: FattyMcButterpants on October 14, 2014, 11:55:05 PM

Quote from: the joint on October 13, 2014, 03:54:34 AM

Quote from: Este Nuno on October 13, 2014, 03:41:31 AM

Quote from: mnmShadyBTC on October 12, 2014, 09:07:06 PM

Quote from: jonald_fyookball on October 12, 2014, 04:43:56 PM

Quote from: mnmShadyBTC on October 12, 2014, 04:41:37 PM

Quote from: leopard2 on October 12, 2014, 02:05:49 PM

Quote from: franky1 on October 11, 2014, 03:10:39 AM

Quote from: Bit_Happy on October 11, 2014, 12:10:53 AM

Over 30 minutes for a single confirmation does not (and will not ever) work in the real world.
(Almost) Every time I start to believe that BTC is "going mainstream", I have a delay waiting for a 30+ minute block.... Shocked

solution:
if your paying for a coffee or a mars bar, dont wait. if your buying a house wait 30 minutes

solution:
if your paying for a coffee or a mars bar, use LTC or DOGE. if your buying a house wait 30 minutes Wink

One confirmation with LTC or DOGE is much less secure then a 0/unconfirmed TX with bitcoin. Both of their network is not secure so it would be easy to attack the network and reverse your TX. Once a TX is propagated throughout the network, nodes will not accept TXs that use the same inputs so you will not be able to broadcast a TX that spends the same inputs, the only way to double spend is to confirm a block that uses at least one of the same inputs as what you just spent. As long as the value of your cup of coffee is less then what it costs to mine the next block then you are almost certainly safe to accept a 0/unconfirmed TX after the TX has been properly propigated

Not disagreeing with you , but isnt LTC fairly secure?

It is the most secure altcoin out there as of now but the short block time gives a misleading picture as to how secure it's network is. Even if it has the same difficulty as bitcoin it's 2.5 minute block time means that it takes 25% of the work to find one block and there will be at least 4x as many orphaned blocks.

If you want to do an apple to apples comparison of network security the LTC network has a difficulty of ~36.5 thousands now, compared to 35 million for bitcoin. Granted BTC and LTC miners are not interchangeable but I think this gives a pretty good comparison as to how easy it would be to attack either network.

The thing is that one confirmation is more secure than zero confirmations as it prevents a whole category of zero confirmation double spend attacks. So that confirmation comparison you're making is only accurate for a comparison of attacks where a rogue miner is actually publishing blocks. And the amount of time you need to wait for the same amount of security is the same so they're actually equivalent. Like I said up thread, the only disadvantage to shorter block times seems to be more orphans, but since it's something that benefits the end user I would think most people who aren't miners would actually be behind such a change. And the smart ones who are miners will see that despite the fact that they will be wasting some hash in some cases, the potential gains in demand and positive perception due to people not having to wait around 30 minutes for a confirmation often enough should make up for it in the long run. That's just conjecture of course, but considering the amount of complaints we hear about Bitcoin's slow block targets I wouldn't be surprised if that was the case.

It's not just more secure, it's infinitely more secure.

Fast-confirm coins absolutely have a niche, in my opinion.

Fast confirm coins are much less secure and having one confirmation on a fast confirm network is less secure then a 0/unconfirmed TX on a 10 minute network. The reason for this is that nodes will not accept a double spend transaction when there is a propagated TX that spends certain inputs. As long as nodes are sufficiently diverse and TXs are able to propagate quickly, then it will not be possible to double spend certain inputs you have spent on a 0/unconfirmed TX unless you find a block and include the double spend TX in that found block. At the very least you will need to wait for enough confirmations to equal one "10 minute" confirmations to get the same security of a 0/unconfirmed TX assuming the hashrate is the same.

The problem with this is that shorter confirmation blockchains also give miners more of an incentive to mine on shorter blockchains to prevent their chain from being orphaned. This will result in more orphaned chains and longer chains, ones that will potentially exceed the equivalent of a 10 minute confirmation

But it is possible to double spend unconfirmed transactions if you know how to do it correctly(not saying it's easy). Which is why the 1 confirmation is better, since that whole category of double spend attacks that are possible are protected against with one confirm.

I think at this point it's pretty clear that Bitcoin could upgrade to 2 minute blocks on the next hard fork and people would be happy with the change. One confirmation, even if it's 1/5th as secure as a one 10 minute confirm is still very secure on the Bitcoin network. I think it would beneficial for Bitcoin to make this change.

Here's Peter Todd's take on unconfirmed transactions: http://www.reddit.com/r/Bitcoin/comments/239bj1/doublespending_unconfirmed_transactions_is_a_lot/

Note: I'm not saying that unconfirmed transactions aren't fine for microtransactions and such.

Changing to 2 minute blocks would not be an upgrade. The only way to double spend an unconfirmed transaction if you do not control the miner that finds a block that confirms a double spend TX is to control a well connected node that miners connect to (you would probably actually need to control several well connected nodes)

Dabs

legendary

Activity: 3416

Merit: 1912

The Concierge of Crypto

I found this from a much older thread.

Quote from: DeathAndTaxes on May 03, 2012, 10:31:49 AM

The 0-confirm double spend is something which has been known and studies for some time.

That being said a 10 sec window for real world face to face transactions is just asinine.

Counter experiment.
Go to starbucks, order, swipe your credit card.
START STOPWATCH
wait for coffee ... wait .... wait ... get coffee ... walk to door ... get in car ... drive out of line of site.
STOP STOP WATCH

I am 99% certain the elapsed time is going to be >10 sec.

There are methods to reduce (but likely never eliminate) the fraud risk of 0-confirm double spends. That being said no payment system is without risk.

Cash: counterfeit bills, theft, bank processing fees (which is why stores "generously" allow cashback)
Credit cards: stolen cards, friendly fraud, never ending cut to VISA/MC.

The goal isn't to reduce fraud/theft/loss to 0%. Doing that likely will drive away consumers. The goal is to reduce the COST OF PROCESSING transactions which includes everything from employee training to fraud to bank fees to infrastructure/logistics.

Este Nuno

legendary

Activity: 826

Merit: 1002

amarha

Quote from: OrientA on October 14, 2014, 03:42:28 AM

Quote from: Este Nuno on October 13, 2014, 06:16:31 AM

So I wonder why people haven't been asking the core team to change the confirm time on the next hard fork? This is a perfect time to do it considering we're very likely going to be doing a hard fork any way to increase the block size from 1 mb.

It takes about 1 hour to make sure the confirmation is solid due to orphan blocks or some kind of attacks. 6 confirmation of 10 minute block takes 1 hour. If you reduce the block time to 1 min, then it will take 60 confirmation to avoid double spend.

Right, so we might as well have shorter confirms since no one is going to be rewriting blocks on the Bitcoin chain for any transaction that's worth less than the cost to do so. And for the ones that are, we wait an hour. Sounds pretty good to me.

Quote from: FattyMcButterpants on October 14, 2014, 11:55:05 PM

Quote from: the joint on October 13, 2014, 03:54:34 AM

Quote from: Este Nuno on October 13, 2014, 03:41:31 AM

Quote from: mnmShadyBTC on October 12, 2014, 09:07:06 PM

Quote from: jonald_fyookball on October 12, 2014, 04:43:56 PM

Quote from: mnmShadyBTC on October 12, 2014, 04:41:37 PM

Quote from: leopard2 on October 12, 2014, 02:05:49 PM

Quote from: franky1 on October 11, 2014, 03:10:39 AM

Quote from: Bit_Happy on October 11, 2014, 12:10:53 AM

Over 30 minutes for a single confirmation does not (and will not ever) work in the real world.
(Almost) Every time I start to believe that BTC is "going mainstream", I have a delay waiting for a 30+ minute block.... Shocked

solution:
if your paying for a coffee or a mars bar, dont wait. if your buying a house wait 30 minutes

solution:
if your paying for a coffee or a mars bar, use LTC or DOGE. if your buying a house wait 30 minutes Wink

One confirmation with LTC or DOGE is much less secure then a 0/unconfirmed TX with bitcoin. Both of their network is not secure so it would be easy to attack the network and reverse your TX. Once a TX is propagated throughout the network, nodes will not accept TXs that use the same inputs so you will not be able to broadcast a TX that spends the same inputs, the only way to double spend is to confirm a block that uses at least one of the same inputs as what you just spent. As long as the value of your cup of coffee is less then what it costs to mine the next block then you are almost certainly safe to accept a 0/unconfirmed TX after the TX has been properly propigated

Not disagreeing with you , but isnt LTC fairly secure?

It is the most secure altcoin out there as of now but the short block time gives a misleading picture as to how secure it's network is. Even if it has the same difficulty as bitcoin it's 2.5 minute block time means that it takes 25% of the work to find one block and there will be at least 4x as many orphaned blocks.

If you want to do an apple to apples comparison of network security the LTC network has a difficulty of ~36.5 thousands now, compared to 35 million for bitcoin. Granted BTC and LTC miners are not interchangeable but I think this gives a pretty good comparison as to how easy it would be to attack either network.

The thing is that one confirmation is more secure than zero confirmations as it prevents a whole category of zero confirmation double spend attacks. So that confirmation comparison you're making is only accurate for a comparison of attacks where a rogue miner is actually publishing blocks. And the amount of time you need to wait for the same amount of security is the same so they're actually equivalent. Like I said up thread, the only disadvantage to shorter block times seems to be more orphans, but since it's something that benefits the end user I would think most people who aren't miners would actually be behind such a change. And the smart ones who are miners will see that despite the fact that they will be wasting some hash in some cases, the potential gains in demand and positive perception due to people not having to wait around 30 minutes for a confirmation often enough should make up for it in the long run. That's just conjecture of course, but considering the amount of complaints we hear about Bitcoin's slow block targets I wouldn't be surprised if that was the case.

It's not just more secure, it's infinitely more secure.

Fast-confirm coins absolutely have a niche, in my opinion.

Fast confirm coins are much less secure and having one confirmation on a fast confirm network is less secure then a 0/unconfirmed TX on a 10 minute network. The reason for this is that nodes will not accept a double spend transaction when there is a propagated TX that spends certain inputs. As long as nodes are sufficiently diverse and TXs are able to propagate quickly, then it will not be possible to double spend certain inputs you have spent on a 0/unconfirmed TX unless you find a block and include the double spend TX in that found block. At the very least you will need to wait for enough confirmations to equal one "10 minute" confirmations to get the same security of a 0/unconfirmed TX assuming the hashrate is the same.

The problem with this is that shorter confirmation blockchains also give miners more of an incentive to mine on shorter blockchains to prevent their chain from being orphaned. This will result in more orphaned chains and longer chains, ones that will potentially exceed the equivalent of a 10 minute confirmation

But it is possible to double spend unconfirmed transactions if you know how to do it correctly(not saying it's easy). Which is why the 1 confirmation is better, since that whole category of double spend attacks that are possible are protected against with one confirm.

I think at this point it's pretty clear that Bitcoin could upgrade to 2 minute blocks on the next hard fork and people would be happy with the change. One confirmation, even if it's 1/5th as secure as a one 10 minute confirm is still very secure on the Bitcoin network. I think it would beneficial for Bitcoin to make this change.

Here's Peter Todd's take on unconfirmed transactions: http://www.reddit.com/r/Bitcoin/comments/239bj1/doublespending_unconfirmed_transactions_is_a_lot/

Note: I'm not saying that unconfirmed transactions aren't fine for microtransactions and such.

Dabs

legendary

Activity: 3416

Merit: 1912

The Concierge of Crypto

Bitcoin is good for any purchase that would otherwise require you to write a check. Think of it as a check that is accepted in seconds, approved in minutes, and cleared in an hour.

Topic: Blocks are still taking way too long. - page 2. (Read 5929 times)