> I think Paul has been pretty upfront about the risks of his model. I think he has been rather misleading in his presentation of the risks. He outlines them in a very technical manner, yes, but then goes on to promote them to lay people as if they're no big deal, which is completely misleading. > By your account bitcoin is already insecure then -- it allows anyone can spend outputs that can be claimed by miners. That is completely different. It is disingenuous to say the two are remotely similar. The two situations have little-to-nothing in common. In the present situation, anyone-can-spend outputs are used by probably less than 0.1% of users, and most software doesn't even allow for the possibility. In Drivechain it's *encouraged-by-design*! - Greg -- Please do not email me anything that you are not comfortable also sharing with the NSA. > On Jul 12, 2017, at 12:34 PM, Chris Stewart > wrote: > > Hi Greg, > > The safest way to ensure everyone's protection to make sure *no one can do anything*. Then we will ALL be safe ;). > > >If so, please leave, you are compromising Bitcoin's security. > > Ok, let's calm down. > > >If I design a car that has a button that randomly causes the brakes to give out if pressed, is that a good idea? Can I justify pushing for such a "feature" just because it's "opt-in"? > > It would be more like "should we allow a car on the road if we know statistically that our brakes give out in every 1/100,000,000 cars"? There is security risks with everything in life -- we need to quantify the risk to see if it is worth taking. I think Paul has been pretty upfront about the risks of his model. I think you did a good job of demonstrating it in the email I cited too. > > >It is how *insecure* systems are designed. > > By your account bitcoin is already insecure then -- it allows anyone can spend outputs that can be claimed by miners. > > >Sure, happy to, as soon as I have it written up in detail. > > I look forward to this! > > -Chris > > On Wed, Jul 12, 2017 at 2:24 PM, Tao Effect > wrote: > Dear Chris, > >> I think this is an unfair characterization. You have to opt into using drivechains. > > I have heard this nonsense repeated countless times in order to justify adopting Drivechain. > > This is not how security works. > > A child can "opt-in" to using a loaded gun, but is it a good idea to make it easy for them to do that? > > No. > > This is effectively the same thing Drivechains is doing. > > It is a request to modify the Bitcoin protocol to make it easy for Bitcoin users to give their Bitcoins to miners. > > Does that sound like a good idea to anyone? > > If so, please leave, you are compromising Bitcoin's security. > > Security is about making it difficult to shoot yourself in the face. > > If I design a car that has a button that randomly causes the brakes to give out if pressed, is that a good idea? Can I justify pushing for such a "feature" just because it's "opt-in"? > > No. That is fallacy. > > It is not how secure systems are designed. > > It is how *insecure* systems are designed. > >> Care to share? I'm unaware if there is. > > > Sure, happy to, as soon as I have it written up in detail. > > Kind regards, > Greg Slepak > > -- > Please do not email me anything that you are not comfortable also sharing with the NSA. > >> On Jul 12, 2017, at 12:19 PM, Chris Stewart > wrote: >> >> Hi Greg, >> >> >Here, you admit that the security of the sidechains allows miners to steal bitcoins, something they cannot do currently. >> >> If I put my coins in an anyone can spend output, a miner will take them. They can do this today. I suggest you try it if you don't believe me :-). You have to be more specific with contract types instead of generically talking about 'all contracts ever'. >> >> > Drivechain is an unmistakeable weakening of Bitcoin's security guarantees. This you have not denied. >> >> I think this is an unfair characterization. You have to opt into using drivechains. Other outputs such as P2PKH/Multisig etc are unaffected by a drivechain output. As Pieter Wuille stated earlier in this thread (and Paul has stated all along), drivechain outputs have a different security model than other contracts. Namely they are controlled by miners. I think we can all agree this is unfortunate, but it is the current reality we live in. I look forward to the day we can solve the 'ownership' problem so we can have trustless interoperable blockchains, but that day is not today. >> >> As a reminder, most users will not have to go through the drivechain withdrawal process. Most withdrawals will be done via atomic swaps. >> >> >There is no reason to weaken Bitcoin's security in such a dramatic fashion. Better options are being worked on, they just take time. >> >> Care to share? I'm unaware if there is. >> >> >https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2017-June/014600.html >> >> Everyone should re-read this email though, this is something that could happen. Paul's design makes it so that if this occurs it is *VERY* obvious. I guess we can argue if there is any difference between an obvious robbery vs a hidden robbery, but I think if we have to pick one or the other the choice is clear to me. Other designs (that I'm aware of) for sidechains had attack vectors that weren't so obvious. >> >> -Chris >> >> >> > >