The Holy Grail! I wish I could kiss the author of Bitmessage on his face.

[fellowtraveler]

I should also mention:

1. Cross-server transfers of funds will involve smart contracts verified by the servers on both sides. (Not by the users.) This is only for OT-server to OT-server, but still worth mentioning. The server itself would have to be "in on" the scheme with Jorg -- and an OT server can't forge receipts.

2. Conversion of one currency into another (on markets) is also performed by the server, and thus you don't have to trust the user you're dealing with. The same goes for escrow, for all internal transactions. Bitmessage allows you to find these other users, but the actual transactions are then performed on OT.

[Zangelbert Bingledack]

Now I'm wondering if OpenCoin will try to buy out Monetas.

[gmaxwell]

OT is able to confer properties of a decentralised network ... and Bitcoin is actually a highly "distributed" network (any node to any node), a special case of decentralisation.

This is a very different definition of distributed/decenteralized than is used in some other places.

The other one is:

Visa's network is distributed— there are servers in many datacenters that share the work. But they are all controlled by a central party, they are not decentralized. Likewise, google, any CDN, basically any highly scalable modern service is distributed.

Decenteralized means there no central authority at all. Very few systems are decenteralized, though arguably many system-of-system federations are at least to a degree decentralized.

Most of the capabilities were already latent in OT. Bitmessage is just what solves discovery.
(In fact you could swap it out for any similar solution which solves discovery.)

Right. What does bitmessage accomplish for you that a tor hidden service wouldn't accomplish? Or an IRC channel?

I thought OT federations needed unjammable broadcast networks in order to prove that they weren't performing inflationary double issuances of the assets they published notes for?

[marcus_of_augustus]

OT is able to confer properties of a decentralised network ... and Bitcoin is actually a highly "distributed" network (any node to any node), a special case of decentralisation.

This is a very different definition of distributed/decenteralized than is used in some other places.

The other one is:

Visa's network is distributed— there are servers in many datacenters that share the work. But they are all controlled by a central party, they are not decentralized. Likewise, google, any CDN, basically any highly scalable modern service is distributed.

Decentralized means there no central authority at all. Very few systems are decenteralized, though arguably many system-of-system federations are at least to a degree decentralized.

As far as I'm aware there are no precise mathematical/engineering definitions for centralized and decentralized networks? Obviously any network without a central hub controller is decentralized, can we agree on that at least?

I dont' know how other people use it but I thought the posted diagrams make clear the differences. I have seen node-to-node type networks (rightmost picture) referred to as distributed for a long time (3 decades already maybe?) but also they recently call them "mesh" networks. To add to the confusion, in many real-time networks applications SCADA (supervisory computer and distributed acquisition) were some of the first decentralised systems on the market, alongside their competitors/counterparts the DCS (distributed control systems) which actually were much more 'centralised'.

Perhaps it is time for someone to sort all this out with a good paper tabulating the network topologies nomenclature, fractal measures, scaling properties, etc? You might be up for that?

[fellowtraveler]

Visa's network is distributed— there are servers in many datacenters that share the work. But they are all controlled by a central party, they are not decentralized. Likewise, google, any CDN, basically any highly scalable modern service is distributed.

Decenteralized means there no central authority at all. Very few systems are decenteralized, though arguably many system-of-system federations are at least to a degree decentralized.

The OT servers are definitely not all controlled by a central anything. Each has its own server operator, presumably.

Also, an OT server is not an authority over anything, but would be more accurately termed a "cloud commodity."

The authority should be the user wallet itself, not any provider/server. Similar to Tahoe-LAFS: provider-independent.

Most of the capabilities were already latent in OT. Bitmessage is just what solves discovery.
(In fact you could swap it out for any similar solution which solves discovery.)

Right. What does bitmessage accomplish for you that a tor hidden service wouldn't accomplish? Or an IRC channel?

I have repeatedly said in this thread that most of the new capability here is latent in Open-Transactions itself, and that Bitmessage only solves discovery.

I have also explicitly said that IRC could most likely be swapped in to replace Bitmessage -- and it would probably still work. (I don't see why not.)

In fact the OT servers themselves could even allow users to post and share certain announcements in order to solve this problem.

A Tor hidden service, as you suggest, would probably also work -- though it seems to be the most centralized form of all the suggestions.

I will definitely be coding it to use an abstraction layer so that different options are possible.

I thought OT federations needed unjammable broadcast networks in order to prove that they weren't performing inflationary double issuances of the assets they published notes for?

The purpose of the broadcast as discussed in this thread is purely for discovery purposes, not for auditing. Each OT server does need to make certain audit information available for query (it doesn't have to be stored in any censorship-proof medium.) But that's not the same thing as the discovery process that makes possible these cross-server wires, cross-server trades, and server/legacy bank transfers.

[boonies4u]

Thanks for all the info FellowTraveler... I have one question though, what are "multi-sig voting pools", and how do they work? Do they actually hold onto funds, or do they just sign off on transactions involving them?

[Rassah]

Just want to say that when I see BMOT, I'm my mind I read it almost like бeгeмoт (begemot, e as in west), which is Russian for hippopotamus or behemoth. That second one sounds strangely foretelling 

[caveden]

For people that just want to buy and sell BTC once in a while - as opposite to those who wish to profit from active trading -, what's the advantage in holding fiat-backed colored coins?
Couldn't the same scheme of escrows and everything be used to acquire BTC directly, instead of going through colored coins?

By the way, I think using alternate chains with merged mining is better than using colored coins in the main chain. Colored coins would likely be "tagged satoshis" if I'm not mistaken. To transfer these colored coins around, you'd need actual bitcoins to pay mining fees (and first of all, you'd need a pool that's willing to ignore recent standards on anti-dust). That means that to move your "dollars IOUs" around you'd need bitcoins to pay the transfers fees. It's kind of weird, and will make this system only useful for those willing to acquire bitcoins - and these ones, AFAICT, will not have any other reason to hold these colored coins than active trading.
If instead you use an alternate chain where only fiat-backed coins exist, the mining fees would have to be payed in fiat-backed coins. That's more reasonable, and may even become a strong competitor to legacy fiat transfer methods, without requiring the user to hold bitcoins not even for a second. That could grow more quickly, and extend the advantages of "Bitcoin the payment network" to those who refuse to hold "Bitcoin the currency".
As an extra, as this alternate chain would have no currency creation, we'd already have an experimentation case for fee-only mining subsidy.

A single alternate chain could be created for holding everything related to the "issuance of tokens". It could hold everything you'd be willing to use colored coins for, I suppose.

[fellowtraveler]

Thanks for all the info FellowTraveler... I have one question though, what are "multi-sig voting pools", and how do they work? Do they actually hold onto funds, or do they just sign off on transactions involving them?

Consider the various forms...

----------------------------------------------------------------

GOLD ISSUER

1. The gold issuer issues gold units onto the OT server.
2. Any user who has the currency contract is able to open accounts denominated in those units, and from there, withdraw cash, write cheques, trade on markets, etc.

===> In this example, the OT server cannot forge any receipts. The only possible crime is inflation, but the gold issuer has an incentive to audit the OT server, which prevents inflation.

----------------------------------------------------------------

EURO COLORED COIN ISSUER

1. The Euro issuer issues colored coins.
2. Users have the option to upload these colored coins to OT servers (preferably via voting pools).
2. Any user who has done this is able to open accounts denominated in those units, and from there, withdraw cash, write cheques, trade on markets, etc.

===> In this example, the OT server cannot forge any receipts. The only possible crime is inflation, but the other voting pool members have an incentive to audit the OT server, which prevents inflation.

----------------------------------------------------------------

BITCOIN WITHOUT MULTI-SIG  (not recommended!)

1. The user uploads the BTC or colored coins to the OT server.
2. The OT server then issues the appropriate units to the user.
3. Whenever the user wants to get his BTC or colored coins back out, he sends a signed request to the OT server along with the units, and the server sends his BTC back to him on the blockchain.

===> In this example, the server would have to be trusted not to disappear and steal all the BTC he's holding.

===> The server would also not have any incentivized entity performing audits to prevent inflation.

===> The one benefit is that the OT server cannot forge any of your receipts (so at least the amount he owes you is provable.)

===> This configuration sucks (I do not recommend it) but this is basically what the entire Bitcoin world has been doing, up until this point.

===> This is why people keep getting screwed in the Bitcoin world. The server just disappears with your money, or gets "hacked."

----------------------------------------------------------------

BITCOIN **WITH** MULTI-SIG

1. The user uploads the BTC or colored coins to a list of BTC addresses, instead of to a single address. Each one of the addresses on this list belongs to a member of the voting pool.
2. Each voting pool member is an OT server. Once BTC or colored coins are in the pool, then only an M-out-of-N vote from the servers -- on the blockchain itself -- can retrieve those coins.
3. If the user wants to get his BTC or colored coins back off of the OT server, he sends a bail-out request, which the server countersigns, and then they forward this to the pool members, who verify the signatures and vote ON THE BLOCKCHAIN to release the coins back to the user.
4. Even if the OT server disappears entirely, the user can still submit a recovery request to the other pool members and get their vote, and get the coins recovered.
5. In answer to your question, the coins just sit in the voting pool the whole time while they are being transacted on the OT server. Ideally they will change hands a hundred times, a thousand times, a million times, before being pulled back out of the pool. The whole point is to enable off-chain transactions on transaction servers, with escrow and markets, etc, and to avoid expensive and traceable blockchain transactions except where necessary.

===> The only crime left to the OT server (who cannot forge receipts) is the crime of inflation. However, the other pool members have an incentive to audit each other, which prevents this crime. Therefore the pool itself replaces the "gold issuer" in the original example. See the Open-Transactions auditing doc.

===> You do not have the trust the individual servers, but you DO have to trust the POOL ITSELF. If a hacker were to gain malicious control over a majority, or 8 out of 10, or whatever, of those servers, then he could steal the funds in the pool.

===> This is why I would go even further, and use a wallet GUI that distributes my funds across multiple pools, and/or uses basket currencies to distribute a single currency across multiple issuers / multiple pools.

Ultimately, you cannot eliminate risk entirely, but you can reduce it, and distribute it, and take advantage of separation of powers.

The concept with OT is to create a wallet-centric experience for automating this, to achieve provider-independence. (Like Tahoe-LAFS.)

Couldn't the same scheme of escrows and everything be used to acquire BTC directly, instead of going through colored coins?

Yes, an issuer can directly issue his IOU units onto an OT server, without having to go through colored coins.

But colored coins are important IMO because they allow you to sever this direct link between issuer and transaction server.

That is, you can still have issuers, and you can still have transaction servers, but they are no longer directly connected.

This is important for liability reasons, and it eliminates the issuer's ability to pick and choose transaction servers (meaning that authorities also cannot pressure the issuer to do so) and it also allows the users to buy and sell the colored coins as commodities, similar to BTC itself. Trust me, if you work through the exact differences between those scenarios, you will see why it's better to issue them as colored coins. (But OT will work either way, yes.)

Colored coins are specifically for Dollar or Euro based currencies, or even gold-based currencies. For BTC, on the other hand, you just use BTC instead of colored coins. Then you have no issuer and the BTC itself is the primary currency being exchanged. OT will work either way.

[caveden]

Colored coins are specifically for Dollar or Euro based currencies, or even gold-based currencies. For BTC, on the other hand, you just use BTC instead of colored coins. Then you have no issuer and the BTC itself is the primary currency being exchanged. OT will work either way.

I can understand the advantages of colored coins over OT-IOUs (well, I guess I can, it's mostly the strong censorship-resistance of cryptocurrencies, right?)

But the use of colored-coins as a method of fiat transferring (i.e., as a direct competitor to PayPalUSD or Dwolla or whatever) would be strongly hindered IMHO if one has to pay bitcoin fees in order to transfer these "fiat coins".
I still think a dedicated chain for holding these tokens would be better than coloring coins in the main chain. Don't you agree?

[fellowtraveler]

Colored coins are specifically for Dollar or Euro based currencies, or even gold-based currencies. For BTC, on the other hand, you just use BTC instead of colored coins. Then you have no issuer and the BTC itself is the primary currency being exchanged. OT will work either way.

I can understand the advantages of colored coins over OT-IOUs (well, I guess I can, it's mostly the strong censorship-resistance of cryptocurrencies, right?)

Right. More specific reasons already detailed above.

But the use of colored-coins as a method of fiat transferring (i.e., as a direct competitor to PayPalUSD or Dwolla or whatever) would be strongly hindered IMHO if one has to pay bitcoin fees in order to transfer these "fiat coins".
I still think a dedicated chain for holding these tokens would be better than coloring coins in the main chain. Don't you agree?

I think it's open for debate, and the market will decide, but personally I disagree.

Why?

Because in my concept, you are only very rarely moving the colored coins in-and-out of the pool. Rather, you are trading them INSIDE OT. So your Nym, INSIDE OT, gives some colored coins to another Nym, INSIDE OT, perhaps by way of market trade or escrow agreement, in return for some other BTC or colored coin INSIDE OT, or in return for some legacy funds in real-world cash or in a real-world bank account.

As I've said, I would not want to move coins in/out of the pool for every transaction. Instead, I would keep them in the pool (in OT) for thousands of transactions, only moving them in/out when absolutely necessary.

...And the reason I would store BTC and colored coins on the main chain, instead of some alternate chain, is because I believe the security is best there. Since they will be "just sitting there" for some time, I prefer they sit somewhere safe.

[caveden]

Because in my concept, you are only very rarely moving the colored coins in-and-out of the pool. Rather, you are trading them INSIDE OT. So your Nym, INSIDE OT, gives some colored coins to another Nym, INSIDE OT, perhaps by way of market trade or escrow agreement, in return for some other BTC or colored coin INSIDE OT, or in return for some legacy funds in real-world cash or in a real-world bank account.

But don't you lose the censorship-resistance features of cryptocurrencies once you deposit them on a particular OT-server? Can't the server be killed, freezing all accounts within it?

[marcus_of_augustus]

Colored coins are specifically for Dollar or Euro based currencies, or even gold-based currencies. For BTC, on the other hand, you just use BTC instead of colored coins. Then you have no issuer and the BTC itself is the primary currency being exchanged. OT will work either way.

I can understand the advantages of colored coins over OT-IOUs (well, I guess I can, it's mostly the strong censorship-resistance of cryptocurrencies, right?)

For people that just want to buy and sell BTC once in a while - as opposite to those who wish to profit from active trading -, what's the advantage in holding fiat-backed colored coins?
Couldn't the same scheme of escrows and everything be used to acquire BTC directly, instead of going through colored coins?

You are also probably not considering the advantages of holding provably backed bitcoin OT tokens versus bitcoins themselves ... vouchers, checks, near instant off-chain TX, untraceable blinded cash, recurring payments, plus other stuff. There are payment advantages (i.e. additional monetary properties) to having OT tokens. Think of OT as a decentralised system layered on top of the bitcoin mesh ... and you start to see the bigger picture.

[fellowtraveler]

Because in my concept, you are only very rarely moving the colored coins in-and-out of the pool. Rather, you are trading them INSIDE OT. So your Nym, INSIDE OT, gives some colored coins to another Nym, INSIDE OT, perhaps by way of market trade or escrow agreement, in return for some other BTC or colored coin INSIDE OT, or in return for some legacy funds in real-world cash or in a real-world bank account.

But don't you lose the censorship-resistance features of cryptocurrencies once you deposit them on a particular OT-server? Can't the server be killed, freezing all accounts within it?

I think if you read the above posts I wrote, you will see that no, you don't lose the censorship-resistance features of the cryptocurrencies, because they are not sent directly to the OT server, but rather, are stored in a voting pool on the blockchain, where those coins are recoverable even in the event that the server itself disappears.

Also -- what is your "account" ?

On OT, your "account" is just your last signed receipt, and you and the OT server both have a copy of it. In fact that's all you really need to keep a copy of -- all the older receipts can be discarded and the system can still prove everything.

If the issuer creates IOUs, and the OT server disappears or "freezes," then the issuer can just re-issue the same units onto a different OT server and it can earn the transaction fees instead. (The problem with this is, what if jurisdictional authorities, or criminals, use violence to force the issuer to abandon a specific transaction server? This is why I recommend adding the cryptocurrency layer...)

If you add the cryptocurrency layer, the issuer will create colored coins and release them (Or the miners create BTC and release them.) Either way, when a user uploads those coins into the multi-sig voting pool, they are safe in the pool (the transaction server can't steal them, and the issuer has no power to pick-and-choose which transaction servers are being used -- the users choose.) If your transaction server disappears or "freezes," then your wallet can just submit a recovery request to the other pool members and get their vote on the blockchain, and get the coins back.

What if the entire POOL gets hacked? Then you are in a bad position. Ultimately the only solution is to have a well-designed wallet which distributes risk across multiple currencies and multiple pools. This is where I think everything is going, and that's what I am building.

[caveden]

I think if you read the above posts I wrote, you will see that no, you don't lose the censorship-resistance features of the cryptocurrencies, because they are not sent directly to the OT server, but rather, are stored in a voting pool on the blockchain, where those coins are recoverable even in the event that the server itself disappears.

My apologies for making you repeat yourself. I ignore how these voting pools work. I'm guessing it's a large multi-signature thing (many people together can replace the signature of the server), but definitely I need to study it a lot more.

Thanks.

[fellowtraveler]

I think if you read the above posts I wrote, you will see that no, you don't lose the censorship-resistance features of the cryptocurrencies, because they are not sent directly to the OT server, but rather, are stored in a voting pool on the blockchain, where those coins are recoverable even in the event that the server itself disappears.

My apologies for making you repeat yourself. I ignore how these voting pools works. I'm guessing it's a large multi-signature thing (many people together can replace the signature of the server), but definitely I need to study it a lot more.

Thanks.

It's just a safe place to store coins, ON THE BLOCKCHAIN, so that an OT server can issue units based on those coins, yet without having the ultimate power to disappear with those coins.

And if the OT server can't steal the actual coins externally on the blockchain, and if it can't forge receipts internally in its own system, then the coins are much more safe than if the transaction server was directly holding them (as we see in the Bitcoin community today.) For example, if your coins are stored at MtGox today, then MtGox ultimately has the power to steal them. They probably won't, but the power is there. But with OT servers, I don't want the server to have such powers. I want to demote the server so that it's merely a cloud commodity, and not an authority.

[Sukrim]

=> SEPA transfer: In Europe, bank users can quickly and easily send transfers to each other's accounts through the SEPA API, and verify whether such transfers have been made.

Could you post a link to this API? As far as I understand it, SEPA is just a standard that is used between banks in the SEPA network, not exposed to actual users.

[caveden]

I think he meant an API provided by the bank to its clients.

[Sukrim]

That's also something that is often limited to business accounts or only in certain countries (German banks often have implemented the HBCI standard). My bank for example tries to rather push it's own branded "apps" instead of allowing an API and there is little to no documentation on what business accounts can get/expect API wise.

Banks are incredibly slow and have no real competition anyways to make them move faster.

[lexxus]

That's also something that is often limited to business accounts

The same is in Russia. I guess for private accounts you have to resort to web scrapping.