DIANNA: the IANA Decentralized design concept

[btc_artist]

Just finished reading the paper.  You have some very, very good ideas.  One thing I will mention, is I encourage you to carefully consider the benefits of merged mining with Bitcoin before discarding the idea.

Namespaces-- It might be a good idea to have namespaces and allow for sub-namespaces.  This way you could have one namespace per "service" (Tor, I2P, a new service in direct competition with ICANN addresses, etc), and then each service can have sub-namespaces which would be analogous with the TLDs we currently have with ICANN.

[1714027786]

1714027786

[1714027786]

1714027786

[matthewh3]

Just finished reading the paper.  You have some very, very good ideas.  One thing I will mention, is I encourage you to carefully consider the benefits of merged mining with Bitcoin before discarding the idea.

Namespaces-- It might be a good idea to have namespaces and allow for sub-namespaces.  This way you could have one namespace per "service" (Tor, I2P, a new service in direct competition with ICANN addresses, etc), and then each service can have sub-namespaces which would be analogous with the TLDs we currently have with ICANN.

I think mergerd mining with Litecoin CPU mining would be a better idea then your not competing against anyone for users as there is no merged mining for CPU mining yet.  So you would attract probally all or most Litecoin miners.

[marcus_of_augustus]

So a rebranded NameCoin ... why don't you just make available a patched NameCoin source-code with your perfect free-market fees formula in it and see if you can get the NameCoin network to adopt it naturally?

Your wiki says :

while DIANNA registration fee is defined by free market agreements

... but where in the design document do you describe this "free market agreements mechanism"?

(As noted this is your main point of difference to NameCoin.)

[thomkaufmann]

FYI, Steve from BitPay recently proposed a method for decentralizing the DNS using bitcoin's blockchain instead of an alt chain: https://bitcointalk.org/index.php?topic=63540.0

[pent]

So a rebranded NameCoin ... why don't you just make available a patched NameCoin source-code with your perfect free-market fees formula in it and see if you can get the NameCoin network to adopt it naturally?

This design will actually destroy namecoin chain if applied. It is fundamentally different from namecoin. It is not NameCoin child, but NameCoin brother.

Your wiki says :

while DIANNA registration fee is defined by free market agreements

... but where in the design document do you describe this "free market agreements mechanism"?

(As noted this is your main point of difference to NameCoin.)

Please read my paper before be opposed. Both your questions are described there.

Domain transaction fee will be set by community like in later Bitcoin, when coinbase dropped to zero. Community will decide what lower and higher limit of fees are acceptable by network to get it live:

Also, this design addresses a crucial issue of setting an acceptable
commission for a domain transaction. It is established on a free market.

Free domain transactions (without linking TxC) are also possible, but their
duration will be depend on how altruistic the miners in the system are.
Transactions with unreasonably high commission are also possible. These
will bring about higher DDiff-s. In this case it will be up to the miners
themselves whether or not to include these transactions into a block,
depending on their hardware facilities. Perhaps such transactions will take a
long while to be completed, waiting while the general complexity reaches an
acceptable level.

All in all, compared to Bitcoin, this system will have a correcting factor
related to the computing difficulty. It will be proportionally tied to additional
financial activity regarding the domain transactions, and average domain
transaction commission as well. This latter value, in its turn, will be established
under free market conditions.

Domain transaction commissions will be returned to
proportionate bonuses for the additional processing undertaken.
minersn as It will be the free choice of the miners which will regulate the
proportionate nature of the additional complexity correction. The higher the
bonus, the higher the complexity – and the higher the risk to get nothing.

[pent]

udpated initial post and wiki with current open issues

[pent]

So we have two issues:
1. How to deal with Bitcoin overwhelming hash power
2. How to safely put a block inside distributed network

What if we try to kill both with one shot?

They summarize into one complex task. But the complex, difficult work - this is what network needs.

In addition to hashing work we will give miners a complex task to find a secure place for block inside our distributed network.

So they not only need to find a hash, but also they need to find a secure place in a huge network for solved block.

[phelix]

Having an alternate dns system is an important thing to secure bitcoins future against possible future dns shutdown attacks. even if its only to prevent these shutdowns from happening.

Why do you split forces and reinvent the wheel?

The only valid argument against namecoin is that domains are too cheap. and that is mostly because of the current lack of interest in the project. So just make marketing for namecoin or try to implement your fee algorithms to namecoin.

Namecoin works fine and is secure. I invite you to http://bitcoinX.bit

[pent]

Oh, pleeeeease...

I already said why i dont like namecoin implementation, scroll up, down and go through links. So big holes in design - this can not grow in something significant.

[matthewh3]

I think your find it difficult to find miners as a lot of people have already adopted merged mining on Bitcoin with Namecoin.  If you was to do merged mining with Litecoin CPU miners your not competing for new miners from anywhere and if you approach the Litecoin community and the major pools I'm sure there all be happy to merge your mining with theirs when your ready to start.  This will lower the chance of a 51% attack.

[ThomasV]

Actually, DIANNA operates like NameCoin, but:
- registration fee is returned to miners (not destroyed) and caused additional proportional difficulty overhead to be not free
- registration fee defined by free market agreements, not by system. System can only suggest its value at first stages like bitcoin tx fee
- coins are not used as a domain name carrier

I agree that this fixes namecoin weaknesses.
 - miners should get paid for the use of disk space induced by domain registration
 - how much they get should be fixed by market
 
However, I believe that the most sensible implementation would be to write domain registrations directly into the Bitcoin blockchain, not to define a new blockchain & currency. Once you realize that the value of coins should not be tied to domain names, the logical conclusion is that we do not need a new blockchain. The existing infrastructure provides enough security.

hum, did I say something Gavin does not want to hear?

[pent]

I agree that this fixes namecoin weaknesses.
 - miners should get paid for the use of disk space induced by domain registration
 - how much they get should be fixed by market
 
However, I believe that the most sensible implementation would be to write domain registrations directly into the Bitcoin blockchain, not to define a new blockchain & currency. Once you realize that the value of coins should not be tied to domain names, the logical conclusion is that we do not need a new blockchain. The existing infrastructure provides enough security.

hum, did I say something Gavin does not want to hear?

Why to overwhelm bitcoin block chain with possible terabytes of domain data? What did Satoshi said about this...

[ThomasV]

I agree that this fixes namecoin weaknesses.
 - miners should get paid for the use of disk space induced by domain registration
 - how much they get should be fixed by market
 
However, I believe that the most sensible implementation would be to write domain registrations directly into the Bitcoin blockchain, not to define a new blockchain & currency. Once you realize that the value of coins should not be tied to domain names, the logical conclusion is that we do not need a new blockchain. The existing infrastructure provides enough security.

hum, did I say something Gavin does not want to hear?

Why to overwhelm bitcoin block chain with possible terabytes of domain data? What did Satoshi said about this...

Satoshi does not own the blockchain. The existence of the blockchain, and the possibility to use it for secure storage of key-value pairs are just facts, and developers cannot do anything about it. Some miners could refuse non standard transactions, but others will be happy to take them.

The scalability of the Bitcoin block chain is a true question, and I think developers should focus on it; it is more important and more urgent than multisig contracts.
However, repeating over and over again "that spamming the blockchain is evil" will not fix things.

[pent]

If Bitcoin developers will put block chain somehow in somewhere like DHT and leave only block headers on clients, the dianna can be futher integrated into bitoin block chain. There will be no need for separate block chain.

The primary problem for now is the every client stores all network data locally. This excludes dianna-like systems from bitcoin chain.

Maybe we need to go by bitcoin distributed storage way?

[ThomasV]

If Bitcoin developers will put block chain somehow in somewhere like DHT and leave only block headers on clients, the dianna can be futher integrated into bitoin block chain. There will be no need for separate block chain.

The primary problem for now is the every client stores all network data locally. This excludes dianna-like systems from bitcoin chain.

This is not different. Every node of your dianna system will need to store all the blockchain locally.
Of course, users of your DNS system do not need the blockchain, they just need to talk to a server.
The same is true with Bitcoin and lightweight clients.

[phelix]

Oh, pleeeeease...

I already said why i dont like namecoin implementation, scroll up, down and go through links. So big holes in design - this can not grow in something significant.

I read through all that unstructured information but I find your arguments invalid.

0. Coins are destroyed Proof
1. NetworkFee descreasing for 50 to zero and will be zero in ~80000 block Proof and has no feedback with network activity like difficulty has.
2. Namecoin was designed for free domains and this is its future Proof
3. Developers have to do system intrusion from time to time to avoid collapse or spam hell Proof Proof

0.) so what?
1.) "
2.) "    Domains will never be free because of the price for name_new. Cheap domains were a design choice as Khal explained to you already (see links above).
3.) Your system will need some time to become as stable as namecoin is already today.

4.) Why not improve the current system? Pretty much seems possible.
5.) I repeat: everybody should pull on one string!

[pent]

0. Coins are destroyed Proof
1. NetworkFee descreasing for 50 to zero and will be zero in ~80000 block Proof and has no feedback with network activity like difficulty has.
2. Namecoin was designed for free domains and this is its future Proof
3. Developers have to do system intrusion from time to time to avoid collapse or spam hell Proof Proof

0.) so what?
1.) "
2.) "    Domains will never be free because of the price for name_new. Cheap domains were a design choice as Khal explained to you already (see links above).
3.) Your system will need some time to become as stable as namecoin is already today.

4.) Why not improve the current system? Pretty much seems possible.
5.) I repeat: everybody should pull on one string!

0. Its a crutch.

Verify domain transactions and verify financial transactions is a different work. Domain transactions can be quitely higher in volume than financial ones. The work for their verification is not paid. name_update is free. name_new fee has no feedback with network activity. System is unbalanced and going to put terabytes of free data on clients HDDs.

Well, if you plan to use this, go ahead.

To fix all problems above namecoin chain needs to be destroyed.

[P4man]

Why to overwhelm bitcoin block chain with possible terabytes of domain data? What did Satoshi said about this...

Could you not have some hybrid approach where the dianna data is stored in a separate blockchain, but the blocks are "approved" in the bitcoin blockchain? So you wouldnt store anything in the bitcoin chain, other than a hash of the last known "approved" dianna block.
disclaimer: I have no idea what Im talking about

[pent]

Could you not have some hybrid approach where the dianna data is stored in a separate blockchain, but the blocks are "approved" in the bitcoin blockchain? So you wouldnt store anything in the bitcoin chain, other than a hash of the last known "approved" dianna block.
disclaimer: I have no idea what Im talking about

This will be a big sync problem. Bitcoin chain is quasi-stable and some its part can be rewritten at any time. What to do in this case...

[P4man]

Could you not have some hybrid approach where the dianna data is stored in a separate blockchain, but the blocks are "approved" in the bitcoin blockchain? So you wouldnt store anything in the bitcoin chain, other than a hash of the last known "approved" dianna block.
disclaimer: I have no idea what Im talking about

This will be a big sync problem. Bitcoin chain is quasi-stable and some its part can be rewritten at any time. What to do in this case...

Not sure I understand the problem, but having an external chain (bitcoin in this case) "approve" the dianna chain actually makes it much more flexible. At any point you could decide to switch to another vetting system, say when bitcoin withers away and litecoin becomes the next big thing.

Anyway, like I said, I have no clue what Im talking about, just airing some random thoughts that other people perhaps can turn in to usable ones .