trust

On Friday 7. August 2015 23.53.43 Adam Back wrote:

> On 7 August 2015 at 22:35, Thomas Zander via bitcoin-dev
> > As we concluded in our previous email, the need to run a node is inversely
> > proportional to the ability (or willingness) to trust others.

[]

> > And lets face it, practically everyone trusts others with their money
> > today.

> Bitcoin's very reason for existence is to avoid that need.  For people
> fully happy to trust others with their money, Bitcoin may not be as
> interesting to them.

I'm making this a thread of its own because this is very serious.

The idea that Bitcoins very reason for existence is to avoid trusting anyone but yourself is something I've heard before, and I have to comment because it is a destructive thought. It is very much untrue because we don't live in a black/white world.

If you look at the history of money (500 years is enough) you may know about business being done in the late 1600s in Europe that included essentially a general ledger that every merchant used and had his own copy of (at least their own bits). Merchant in France used a system that when they bought stock from one company they didn't give them money, they instead gave them a IOU-style piece of paper. To break your promise meant to be evicted from their money system. Which to a merchant in that time is equal to starvation.

The point was NOT to trust no-one, the point was to trust everyone, but keep everyone honest by keeping the ledger open and publicly available.

Bitcoins current model to decentralize and distribute trust has historical precedent and is known to work. It was abandoned when Newton started the mint in London because that allowed international trade. And their system didn't scale.

On a tangent; What we saw with the Internet is growth because of a lack of centralized controller. This does not mean lack of trust in your neighbours. Internet grew because permissionless innovation was allowed. Not by going from one extreme of central trust to the other extreme of no trust. It flourished just by stepping out of the trust-one party extreme.

What Adam Black and probably some others must understand is that there is a whole spectrum between having a monopoly on trust and every player having their own node.

Bitcoin sole reason for existence is because it is the first every system that has global reach and does not need a central trusted party. It is, in other words, the first alternative that for the very first time in centuries that allows innovation without permission.

> For people
> fully happy to trust others with their money, Bitcoin may not be as
> interesting to them.

So, this is to black/white. And also wrong.

This thinking will block growth towards the thing you want, and leave you without any toys at all. For instance it is perfectly all right to have a central player in a poor country that helps millions of unbanked to use Bitcoin as their first international payment system.

I can only try to convince you to change your worldview be explaining some history and concepts you may have missed, if you don't thats fine with me. I do, however, have to ask you to assume people will not like Bitcoin and will not use it because they don't fit your worldview. That will ultimately hurt billions of people.

If you are saying that some people are happy trusting other people, and so would be perfectly fine with off-chain use of Bitcoin, then we agree and I already said that off-chain use case would be a constructive thing for someone to improve scale and interoperability of in the post you are replying to. However that use case is not a strong argument for weakening Bitcoin's security to get to more scale for that use case.

In a world where we could have scale and decentralisation, then of course it would be nice to provide people with that outlook more security than they seem to want. And sometimes people dont understand why security is useful until it goes wrong, so it would be a useful thing to do. (Like insurance, your money being seized by paypal out of the blue etc). And indeed providing security at scale maybe possible with lightning like protocols that people are working on.

Adam

I didn't say off-chain, and gave an example of on-chain usecase with trusted middleman.

So, no, that's not what I meant.

Sent on the go, excuse the brevity.
  Original Message
From: Adam Back
Sent: Saturday, 8 August 2015 09:50
To: Thomas Zander
Cc: Bitcoin Dev
Subject: Re: [bitcoin-dev] trust

If you are saying that some people are happy trusting other people, and so would be perfectly fine with off-chain use of Bitcoin, then we agree and I already said that off-chain use case would be a constructive thing for someone to improve scale and interoperability of in the post you are replying to. However that use case is not a strong argument for weakening Bitcoin's security to get to more scale for that use case.

In a world where we could have scale and decentralisation, then of course it would be nice to provide people with that outlook more security than they seem to want. And sometimes people dont understand why security is useful until it goes wrong, so it would be a useful thing to do. (Like insurance, your money being seized by paypal out of the blue etc). And indeed providing security at scale maybe possible with lightning like protocols that people are working on.

Adam

On 08/08/2015 01:10 PM, Thomas Zander via bitcoin-dev wrote:

> On Friday 7. August 2015 23.53.43 Adam Back wrote:
>> On 7 August 2015 at 22:35, Thomas Zander via bitcoin-dev
>>> As we concluded in our previous email, the need to run a node
>>> is inversely proportional to the ability (or willingness) to
>>> trust others.
> []
>>> And lets face it, practically everyone trusts others with their
>>> money today.
>
>> Bitcoin's very reason for existence is to avoid that need.  For
>> people fully happy to trust others with their money, Bitcoin may
>> not be as interesting to them.
>
> I'm making this a thread of its own because this is very serious.
>
> The idea that Bitcoins very reason for existence is to avoid
> trusting anyone but yourself is something I've heard before, and I
> have to comment because it is a destructive thought. It is very
> much untrue because we don't live in a black/white world.

I think the context is not that "Bitcoin's reason for existence is to avoid trusting anyone but yourself", as you state above. This kind of dystopia is not, in my view, what the advocates of trustlessness and security are implying.

There are those instances where we have no option other than to trust a centralized authority. They are mostly self-appointed and do not have the interests of the rest of us in mind. The central bank is the obvious example - they impose their currency as "official" and then devalue our savings and purchasing power through money supply inflation - without our permission and in betrayal of the trust relationship. Bitcoin allows one to hold money and to conduct money transactions, transmit value, and so forth without having to trust *them* - the central bank, or anyone.

In another example, we want to conduct escrow and have to trust the notary, because of his reputation and framed certificate, but Bitcoin multisig makes the degree of trust between the escrow parties irrelevant, so no apparently trustworthy "gentleman" merchant (back then or now) can socially engineer a transaction and scam us.

I don't know if my reply and its examples is over-simplistic but it seems you were making a moral appeal that the notion of trustlessness was destructive - I just wanted to contextualize it to relevant use cases. It follows, and this is what I understand from Adam's message, that security and protection of decentralization are paramount concerns if we want to retain the trustlessness that makes Bitcoin so useful and powerful.

>> The point was NOT to trust no-one, the point was to trust everyone, but

keep everyone honest by keeping the ledger open and publicly available.

Trust takes many different forms and is not a binary function. You trust a surgeon to do an operation and a pilot to fly a jet, but not vice versa. To trust someone explicitly, you need to know who they are. Most social structures work without explicit identity and they still function quite well. For example companies are mostly anonymous to the consumer - if you buy something in a shop you trust a chain of people producing that good. A priori there is little reason to trust others, but rather that trust is already developed through social institutions. Money is one such institution with specific trust problems, and the history of money is indeed a very good way to study these problems. Unfortunately in Bitcoin development such insights are rare to find.

Lightning assumes explicit trust and ID - much like Ripple. That's not going to work, and I'm surprised that someone with basic knowledge of crypto doesn't see this problem. Having explicit counter-parties is something very different from Bitcoin where the entity doing transactions verification is unknowable and changes all the time. Users of Bitcoin trust nodes doing the verification because they know it is in their best interest to be honest. Neither Sidechains nor LT have preserve that important property, and so IMO there are no good proposals to make Bitcoin scale (if that is possible at all).

On Sat, Aug 8, 2015 at 10:54 AM, Thomas Zander via bitcoin-dev < bitcoin-dev@lists•linuxfoundation•org> wrote:

> I didn't say off-chain, and gave an example of on-chain usecase with
> trusted middleman.
>
> So, no, that's not what I meant.
>
> Sent on the go, excuse the brevity.
>   Original Message
> From: Adam Back
> Sent: Saturday, 8 August 2015 09:50
> To: Thomas Zander
> Cc: Bitcoin Dev
> Subject: Re: [bitcoin-dev] trust
>
> If you are saying that some people are happy trusting other people,
> and so would be perfectly fine with off-chain use of Bitcoin, then we
> agree and I already said that off-chain use case would be a
> constructive thing for someone to improve scale and interoperability
> of in the post you are replying to. However that use case is not a
> strong argument for weakening Bitcoin's security to get to more scale
> for that use case.
>
> In a world where we could have scale and decentralisation, then of
> course it would be nice to provide people with that outlook more
> security than they seem to want. And sometimes people dont understand
> why security is useful until it goes wrong, so it would be a useful
> thing to do. (Like insurance, your money being seized by paypal out
> of the blue etc). And indeed providing security at scale maybe
> possible with lightning like protocols that people are working on.
>
> Adam
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists•linuxfoundation•org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>

Interesting point of view Thomas! I agree that if we only think towards one single direction (treat trust as a super bad thing) we might miss some good features (or scalability levels) among the way.

Benjamin:

> Lightning assumes explicit trust and ID - much like Ripple. That's
> not going to work, and I'm surprised that someone with basic
> knowledge of crypto doesn't see this problem. Having explicit
> counter-parties is something very different from Bitcoin where the
> entity doing transactions verification is unknowable and changes
> all the time.

Can explain why exactly do you think this? What is the problem that you see in lightning model exactly? I am not arguing, maybe you are right and there is a part of the lightning network proposal which I missed, so that is why I am asking for clarification here.

Lightning doesn't require explicit trust, worst case scenario you can end up with coins blocked until next in-chain broadcast. It depends on each and very hub, obviously there will also be trusted, identified public hubs but we can also have anonymous hubs.

On 8/8/2015 12:24 PM, Benjamin via bitcoin-dev wrote:

>>> The point was NOT to trust no-one, the point was to trust
>>> everyone, but keep everyone honest by keeping the ledger open
>>> and publicly available.
>
> Trust takes many different forms and is not a binary function. You
> trust a surgeon to do an operation and a pilot to fly a jet, but
> not vice versa. To trust someone explicitly, you need to know who
> they are. Most social structures work without explicit identity and
> they still function quite well. For example companies are mostly
> anonymous to the consumer - if you buy something in a shop you
> trust a chain of people producing that good. A priori there is
> little reason to trust others, but rather that trust is already
> developed through social institutions. Money is one such
> institution with specific trust problems, and the history of money
> is indeed a very good way to study these problems. Unfortunately in
> Bitcoin development such insights are rare to find.
>
> Lightning assumes explicit trust and ID - much like Ripple. That's
> not going to work, and I'm surprised that someone with basic
> knowledge of crypto doesn't see this problem. Having explicit
> counter-parties is something very different from Bitcoin where the
> entity doing transactions verification is unknowable and changes
> all the time. Users of Bitcoin trust nodes doing the verification
> because they know it is in their best interest to be honest.
> Neither Sidechains nor LT have preserve that important property,
> and so IMO there are no good proposals to make Bitcoin scale (if
> that is possible at all).
>
> On Sat, Aug 8, 2015 at 10:54 AM, Thomas Zander via bitcoin-dev
> <bitcoin-dev@lists•linuxfoundation•org
> <mailto:bitcoin-dev@lists.linuxfoundation.org>> wrote:
>
> I didn't say off-chain, and gave an example of on-chain usecase
> with trusted middleman.
>
> So, no, that's not what I meant.
>
> Sent on the go, excuse the brevity. Original Message From: Adam
> Back Sent: Saturday, 8 August 2015 09:50 To: Thomas Zander Cc:
> Bitcoin Dev Subject: Re: [bitcoin-dev] trust
>
> If you are saying that some people are happy trusting other
> people, and so would be perfectly fine with off-chain use of
> Bitcoin, then we agree and I already said that off-chain use case
> would be a constructive thing for someone to improve scale and
> interoperability of in the post you are replying to. However that
> use case is not a strong argument for weakening Bitcoin's security
> to get to more scale for that use case.
>
> In a world where we could have scale and decentralisation, then of
> course it would be nice to provide people with that outlook more
> security than they seem to want. And sometimes people dont
> understand why security is useful until it goes wrong, so it would
> be a useful thing to do. (Like insurance, your money being seized
> by paypal out of the blue etc). And indeed providing security at
> scale maybe possible with lightning like protocols that people are
> working on.
>
> Adam

On 8 August 2015 at 09:54, Thomas Zander <thomas@thomaszander•se> wrote:

> I didn't say off-chain, and gave an example of on-chain usecase with trusted middleman.

That's basically the definition of off-chain. When we say MtGox or coinbase etc are off-chain transactions, that is because a middle man has the private keys to the coins and gave you an IOU of some kind.

Adam

> Trust takes many different forms and is not a binary function.

Many Bitcoiners have a rather unusual notion of trust. While many state the established financial systems cannot be trusted they imply that many within the Bitcoin world need to be trusted. There are some very irresponsible and unusual people who have authority to do things in the Bitcoin world such as manage the Wiki, the Github repository, and are given an emergency alert key. Some of these people are only known by some sort of screen name. If anyone questions any of these processes they are chastised and labeled a "troll". For instance, anyone who questioned Mt. Gox back in 2013 was labeled a "buttcoiner" or something similar. So I would say many people, including many on this list, don't fully understand the claims they are making about the "truslessness" of Bitcoin.

 >Lightning doesn't require explicit trust,

There seems to be semantic issues with describing trust and the
lightning network.  You need to trust the other party during the
transactions in the channel for it to work.  Yes there is way to cash
out if the other party does not cooperate but that is not the same thing
as saying it is trustless.  Many people are confused about this part of
decentralization where the network has to take time to reach consensus.
  Some people think that you can use some kind of technological trick to
avoid this lag time and still have a completely trustless system.  It is
a tradeoff.

Russ

>>  What is the problem that

you see in lightning model exactly?

How do you know who is who online? If Alice and Bob want to transact and haven't exchanged keys before they need public-key infrastructure out-of-band to identify themselves. Which means they are using SSL and Certificate authorities and trust them. If you have non-cooperative hubs they could flood the network and make it unusable. And why should hubs cooperate? There are no incentives in the system.

On Sat, Aug 8, 2015 at 1:08 PM, s7r <s7r@sky-ip•org> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
>
> Interesting point of view Thomas! I agree that if we only think
> towards one single direction (treat trust as a super bad thing) we
> might miss some good features (or scalability levels) among the way.
>
> Benjamin:
> > Lightning assumes explicit trust and ID - much like Ripple. That's
> > not going to work, and I'm surprised that someone with basic
> > knowledge of crypto doesn't see this problem. Having explicit
> > counter-parties is something very different from Bitcoin where the
> > entity doing transactions verification is unknowable and changes
> > all the time.
>
> Can explain why exactly do you think this? What is the problem that
> you see in lightning model exactly? I am not arguing, maybe you are
> right and there is a part of the lightning network proposal which I
> missed, so that is why I am asking for clarification here.
>
> Lightning doesn't require explicit trust, worst case scenario you can
> end up with coins blocked until next in-chain broadcast. It depends on
> each and very hub, obviously there will also be trusted, identified
> public hubs but we can also have anonymous hubs.
>
> On 8/8/2015 12:24 PM, Benjamin via bitcoin-dev wrote:
> >>> The point was NOT to trust no-one, the point was to trust
> >>> everyone, but keep everyone honest by keeping the ledger open
> >>> and publicly available.
> >
> > Trust takes many different forms and is not a binary function. You
> > trust a surgeon to do an operation and a pilot to fly a jet, but
> > not vice versa. To trust someone explicitly, you need to know who
> > they are. Most social structures work without explicit identity and
> > they still function quite well. For example companies are mostly
> > anonymous to the consumer - if you buy something in a shop you
> > trust a chain of people producing that good. A priori there is
> > little reason to trust others, but rather that trust is already
> > developed through social institutions. Money is one such
> > institution with specific trust problems, and the history of money
> > is indeed a very good way to study these problems. Unfortunately in
> > Bitcoin development such insights are rare to find.
> >
> > Lightning assumes explicit trust and ID - much like Ripple. That's
> > not going to work, and I'm surprised that someone with basic
> > knowledge of crypto doesn't see this problem. Having explicit
> > counter-parties is something very different from Bitcoin where the
> > entity doing transactions verification is unknowable and changes
> > all the time. Users of Bitcoin trust nodes doing the verification
> > because they know it is in their best interest to be honest.
> > Neither Sidechains nor LT have preserve that important property,
> > and so IMO there are no good proposals to make Bitcoin scale (if
> > that is possible at all).
> >
> > On Sat, Aug 8, 2015 at 10:54 AM, Thomas Zander via bitcoin-dev
> > <bitcoin-dev@lists•linuxfoundation•org
> > <mailto:bitcoin-dev@lists.linuxfoundation.org>> wrote:
> >
> > I didn't say off-chain, and gave an example of on-chain usecase
> > with trusted middleman.
> >
> > So, no, that's not what I meant.
> >
> > Sent on the go, excuse the brevity. Original Message From: Adam
> > Back Sent: Saturday, 8 August 2015 09:50 To: Thomas Zander Cc:
> > Bitcoin Dev Subject: Re: [bitcoin-dev] trust
> >
> > If you are saying that some people are happy trusting other
> > people, and so would be perfectly fine with off-chain use of
> > Bitcoin, then we agree and I already said that off-chain use case
> > would be a constructive thing for someone to improve scale and
> > interoperability of in the post you are replying to. However that
> > use case is not a strong argument for weakening Bitcoin's security
> > to get to more scale for that use case.
> >
> > In a world where we could have scale and decentralisation, then of
> > course it would be nice to provide people with that outlook more
> > security than they seem to want. And sometimes people dont
> > understand why security is useful until it goes wrong, so it would
> > be a useful thing to do. (Like insurance, your money being seized
> > by paypal out of the blue etc). And indeed providing security at
> > scale maybe possible with lightning like protocols that people are
> > working on.
> >
> > Adam
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.22 (MingW32)
>
> iQEcBAEBCAAGBQJVxeMaAAoJEIN/pSyBJlsRJFoH/RbgArUMJStQwF92XZk99dUd
> 0xI/VU1goFLDFiFVkrea7uNWUrWw0GM9nDq0kTIV+mTi9rTYgWKlgA1XZnPusr35
> GpDhXxoG3mJmay9AX1fezrZjGmCZPCjSnPWa+BeQCSMXnVchZX0U4XZSwgD7qTIU
> 7o4r5JIDuGxXyPcwECnB7ePmZ8xA2QGQaMW6nnMhlA4KCanSd5/78kcpUp/kGAJ1
> chjhV2g7tAeq3NMs2IXeIMiEAqji0B7RRAejviBg9CAwbpo4dP3dRz8hv/qPx6K0
> Mu6jHczCQOUyAHagwG8q4+laMbkskVETw18NwluspOZi3inxvVpOD60CDqSZPS4=
> =ogMZ
> -----END PGP SIGNATURE-----
>

On Saturday 8. August 2015 12.54.36 Adam Back wrote:

> On 8 August 2015 at 09:54, Thomas Zander <thomas@thomaszander•se> wrote:
> > I didn't say off-chain, and gave an example of on-chain usecase with
> > trusted middleman.
> That's basically the definition of off-chain.  When we say MtGox or
> coinbase etc are off-chain transactions, that is because a middle man
> has the private keys to the coins and gave you an IOU of some kind.

I'm so sorry to have to correct you again, and please don't feel bad about misreading my post twice.

Sending something to another Bitcoin on-chain user is really on-chain. Please believe me when I say that I actually understand my own example.

On Sat, Aug 8, 2015 at 2:37 PM, Thomas Zander via bitcoin-dev <bitcoin-dev@lists•linuxfoundation•org> wrote:

> On Saturday 8. August 2015 12.54.36 Adam Back wrote:
>> On 8 August 2015 at 09:54, Thomas Zander <thomas@thomaszander•se> wrote:
>> > I didn't say off-chain, and gave an example of on-chain usecase with
>> > trusted middleman.
>> That's basically the definition of off-chain.  When we say MtGox or
>> coinbase etc are off-chain transactions, that is because a middle man
>> has the private keys to the coins and gave you an IOU of some kind.
>
> I'm so sorry to have to correct you again, and please don't feel bad about
> misreading my post twice.
>
> Sending something to another Bitcoin on-chain user is really on-chain. Please
> believe me when I say that I actually understand my own example.

Adam, I think he means a multisig escrow transaction where the escrow is trusted by both parties, and other examples like that. But I don't see how that is relevant, allowing trust to be involved in different ways is a feature, but it's optional. I think the point "you don't need to trust anyone to use Bitcoin" remains.

> I think the point "you don't need to trust anyone to use Bitcoin" remains.

I don't think that is a true statement. Users need to trust the mining system is working as intended. Users also need to trust the developers to a certain extent. It is about levels of trust and how much you need to trust these things. When someone says something is "trustless" or "decentralized" they are really saying that it reaches some sort of "acceptable" level in their mind to "brand" it that way.

Russ

On Monday 10. August 2015 22.17.52 Jorge Timón wrote:

> But I don't see how that is relevant, allowing trust to be involved in
> different ways is a feature, but it's optional.

Agreed.

> I think the point "you don't need to trust anyone to use Bitcoin" remains.

yes, and thats fine.

The argument that Adam was making was the other extreme; that Bitcoin was useless the moment he has to trust any 3rd party. And that is why I started this thread, because that idea that Bitcoin looses its functionality when you deviate from the trust-no-one path, is destructive.

For instance I suggested a week ago that the Chinese firewall (blocksize propagation) problem could maybe be solved by having a chinese miner partially trust some full nodes on Internet hubs. For instance in Amsterdam, one in Stockholm, etc. And then propagation of a newly mined block would make the miner only sent the header to this server where a full x-MB block is then created by some specialized software that bases it on the mempool of its bitcoind.

Propagation of a Chinese miner then suddenly doesn't care about blocksize when hopping over the chinese firewall. You'd only send a small amount of data, likely around 20Kb for even 8Mb blocks.

A critic would argue its a useless strategy because you'd have to trust the data center operators. I'd argue that its a risk, for sure, but one that can be mitigated easily by having various datacenters around the world where you'd run your software so you'd be able to check the validity of each.

Risks can only be assessed fairly if people are less black/white in their thinking.

I love that about Bitcoin, it combines technical specialized thinking with Economics and planning. If you don't have both, you probably end up rejecting the best ideas.

> Adam, I think he means a multisig escrow transaction where the escrow
> is trusted by both parties, and other examples like that.

Yes, thats one example. Thanks Jorge!

A similar but relevant example would be like the M-Pesa example. People that have a phone but not a smart phone can choose to have their private keys stored at a trusted company in their own country. Some payments can be off-chain, but as soon as you deal with people outside of this small community they would be on-chain.

When we are talking about remittances, this means one of the two parties does not have an account at the trusted party and as such this will be an on-chain transaction. Same for a farmer in Nairobi buying equipment from a company in Cairo, or a Internet startup in Kampala/Uganda selling services to users in Europe and those users send their payments on-chain.

All of these examples are about the largest group of people in the world; the unbanked. Bitcoin could mean a huge change to these people and I care a lot about this usecase. We need bigger bocks for these usecases as LN will not do anything for them.

On Sat, Aug 8, 2015 at 6:10 AM, Thomas Zander via bitcoin-dev <bitcoin-dev@lists•linuxfoundation•org> wrote:

> The idea that Bitcoins very reason for existence is to avoid trusting anyone
> but yourself is something I've heard before, and I have to comment because it
> is a destructive thought. It is very much untrue because we don't live in a
> black/white world.

[...]

> The point was NOT to trust no-one, the point was to trust everyone, but keep
> everyone honest by keeping the ledger open and publicly available.

I think you are drawing a disction that does not exist; rather it's a disagreement about terminology.

If a system exists that provides high confidence of faithful behavior even from malicious parties, then it's one this community would call "trustless" (assuming the security properties are strong enough). The result is a system where it's possible to trust everyone.

Trust in this case has multiple meanings. In one meaning trust is an (well founded) expectation of faithful behavior. In another it is a blind reliance. When "trust everyone" is used, it's speaking to the first definition, when the trustless definition is used it's referring to the second defintion-- without blind faith.

A trustless (second def) system allows its users to trust (first def) everyone, even the inherently untrustworthy (second def). In doing so, the considerable cost and inequality created by the maintaince of trust (second definition) relations is mitigated, and the availablity of faithful performance increased. Doing so is a prerequsite to having a strongly decenteralized system, because otherwise trust requiremets drive the enviroment towards natural monopolies (as it's cheaper and more effective for more people to trust (second def) a smaller number of parties.

Less philosophically, if you're willing to have systems defined by trust (first definition) (e.g. you do not believe that what I descibed above coveys value, or hope that witl a small number of very trusted parties external factors will transform blind faith into a rational expectation of faithful performance) then there are _much_ more technically superior ways to structure a system than Bitcoin does today-- ones that acheive much greater performance, flexibility, reliablity, and better security (ignoring any insecurity that arises as a result of the trusted parties strong assumptions).

If one wants to layer trust based things on top of a trust mitigating framework, they can do so-- and enjoy efficiencies from doing so. Doing the converse doesn't appear really possible.

Hi Benjamin,

On Sat, Aug 08, 2015 at 02:01:58PM +0200, Benjamin via bitcoin-dev wrote:

> How do you know who is who online?

If a node is not online, then the payment can be cancelled and re-routed.

> If Alice and Bob want to transact and haven't exchanged keys before
> they need public-key infrastructure out-of-band to identify
> themselves. Which means they are using SSL and Certificate authorities
> and trust them.

Lightning doesn't solve the key exchange problem (perhaps something like Namecoin will help in the future). Bitcoin faces this problem today. How do you know the bitcoin address belongs to the recipient without trusting CAs? What if, in the case of the majority of bitcoin payments today, the bitcoin address was not signed and the recipient claimed to have never received their funds? There should be signed proof of payment in every transaction for this reason.

> If you have non-cooperative hubs they could flood the network and make
> it unusable. And why should hubs cooperate? There are no incentives in
> the system.

There are some incentives towards keeping the system functional via fees. If you attempt to flood the system, you'll likely be paying some fees -- someone running a node will not interpret it as an attack, as they're getting some money (probabably substantially higher as they will increase fees to ensure network availability).

I agree that it's very important to think through varius attack models.