Blockchain vs DAG (Byteball's concencus algorithm).

[matein30]

Hi,

DAG (directed acyclic graph) seems to good to be true, no need for miners. No block size issue, no %51 attack. Finalization. More decentrilized.
What are the cons of DAG vs blockchain?
Can bitcoin somehow adapt DAG?

Thanks for answers.

[thejaytiesto]

Hi,

DAG (directed acyclic graph) seems to good to be true, no need for miners. No block size issue, no %51 attack. Finalization. More decentrilized.
What are the cons of DAG vs blockchain?
Can bitcoin somehow adapt DAG?

Thanks for answers.

Good thread. I was also wondering this myself. If this was the case, wouldn't this coin be worth more than BTC already? there must be something, there are always tradeoffs.

I remember reading iamnotback commenting on some Byteball flaws but I dont remember the details.

I don't own any, would like to know more to see if it's worth holding or not.

[Come-from-Beyond]

Can bitcoin somehow adapt DAG?

No. Bitcoiners are scared of changes.

[matein30]

If this was the case, wouldn't this coin be worth more than BTC already? there must be something, there are always tradeoffs.

Coin has distribution problem right now. 87% of all coins are at developer's possession. He tries to distribute it fairly with bitcoin binding. But it will take some time. This why this coin is worthless now.

[cryptohunter]

Hi,

DAG (directed acyclic graph) seems to good to be true, no need for miners. No block size issue, no %51 attack. Finalization. More decentrilized.
What are the cons of DAG vs blockchain?
Can bitcoin somehow adapt DAG?

Thanks for answers.

Good thread. I was also wondering this myself. If this was the case, wouldn't this coin be worth more than BTC already? there must be something, there are always tradeoffs.

I remember reading iamnotback commenting on some Byteball flaws but I dont remember the details.

I don't own any, would like to know more to see if it's worth holding or not.

As I remember it he said BB was one of the only coins he said he'd found interesting to study on here and that that was very rare. Whether he said it was superior in any way to bitcoin I am not certain. If you go to the BB thread there is quite a bit of discussion between him and tonych. Tech wise folks may gain some details there.

[coldmoldy]

IOTA is the best DAG out right now. Zero transaction fees unlike ByteBall.

And they recently partnered with Ubuntu: https://insights.ubuntu.com/2017/02/20/iota-iot-revolutionized-with-a-ledger/

And Innogy: https://innovationhub.innogy.com/news-event/3lj6DX6LS0C4SyKmgEOiQq/meet-our-future-customers--machines-with-wallets


What other coin has done this?

[Big Naturals]

How many current projects using DAG?

[CryptKeeper]

IOTA is the best DAG out right now. Zero transaction fees unlike ByteBall.

And they recently partnered with Ubuntu: https://insights.ubuntu.com/2017/02/20/iota-iot-revolutionized-with-a-ledger/

And Innogy: https://innovationhub.innogy.com/news-event/3lj6DX6LS0C4SyKmgEOiQq/meet-our-future-customers--machines-with-wallets


What other coin has done this?

Yes, IOTA is very innovative and seems to be a good choice for IOT (internet of things, machine2machine communication).

Byteball is the best choice for human2human communication and transfer of value. An easy to use interface and lots of features for real-world use cases (trading bot, merchant bot, assets, private coins, chat).

IMHO these projects are complementary to each other, both have a great future! 

[iamnotback]

I have an entire section (#8) of my whitepaper dedicated to analyzing and explaining Byteball.

It is an interesting discovery. It has flaws (and not just related to the consensus algorithm). And my design is not Byteball, but there is one aspect which is similar to one aspect of Byteball, yet my design would not work if it was only Byteball.

One of the main flaws of Byteball that you should be aware of besides the messed up transaction fees and distribution, is that the consensus of the blockchain can become entirely stuck if > 50% of the witnesses collude or stop functioning. Then it requires a (political gridlock potentially) hardfork to unstuck the blockchain. Ditto Tendermint, Casper, and all Byzantine agreement variants (that includes Bitshares/Graphene/Steem's DPoS, yet the whales can vote to unstuck it without a hardfork). Note there is a mechanism for voting in new witnesses, but because it requires a total order so it will in practice never work out because total orders don't exist in nature.

Yeah Byteball has something important in it, but it is not really the solution.

Iota has an entire different consensus than Byteball, and my stance is it won't work without centralized servers. We've already been down that technical debate in the past, so I won't repeat it and get trolled again by Iota shills. Readers can either learn to recognize that I know what I say, or they can suffer the same fate as all the other times I've been doubted.

And no, Bitcoin can't adopt any of these things. Bitcoin isn't ever going to be anything more than what it already is (which is sufficient). Learn to accept that and move on. Lightning Networks is never going to work (decentralized, which is thus the same as saying it is never going to scale up in terms of relevance).

[Come-from-Beyond]

Iota has an entire different consensus than Byteball, and it doesn't work without centralized servers.

Do you know Java? Can you demonstrate how consensus of IOTA without centralized servers breaks once https://github.com/Come-from-Beyond/Tangle simulator is ready?

[CyanFox]

Can bitcoin somehow adapt DAG?

No. Bitcoiners are scared of changes.

Not only changes, it should be: miners do not like anything let them get less profit, so they embrace BU and don't like SegWit.

[escapefrom3dom]

Can bitcoin somehow adapt DAG?

No. Bitcoiners are scared of changes.

Not only changes, it should be: miners do not like anything let them get less profit, so they embrace BU and don't like SegWit.

they are going to expect less profit in the future so why not to became byteball's witnesses. 

[Rando]

SPECTRE Protocol: DAG Structure to Solve Scalability
http://www.cs.huji.ac.il/~yoni_sompo/pubs/16/SPECTRE_complete.pdf

[Come-from-Beyond]

Iota has an entire different consensus than Byteball, and it doesn't work without centralized servers.

Do you know Java? Can you demonstrate how consensus of IOTA without centralized servers breaks once https://github.com/Come-from-Beyond/Tangle simulator is ready?

I've updated the repo. With Completely Random Walker without attackers global consensus on 84.5% of all transactions was achieved. So the claim "it doesn't work without centralized servers" has already been refuted. The next step is to add Random-Walk Monte Carlo algorithm to show that even under attacks the consensus on significant portion of transactions (subject to connectivity between clusters) is still achieved.

[SatoNatomato]

Iota has an entire different consensus than Byteball, and it doesn't work without centralized servers.

Do you know Java? Can you demonstrate how consensus of IOTA without centralized servers breaks once https://github.com/Come-from-Beyond/Tangle simulator is ready?

I've updated the repo. With Completely Random Walker without attackers global consensus on 84.5% of all transactions was achieved. So the claim "it doesn't work without centralized servers" has already been refuted. The next step is to add Random-Walk Monte Carlo algorithm to show that even under attacks the consensus on significant portion of transactions (subject to connectivity between clusters) is still achieved.

Is this a joke?

You think people will allow machines to spend/receive money and rely on only a probability that their transactions will succeed? And rely on another probability that they have the "global" tip and are not sitting in a cluster? Why not make this simpler and just roll a dice and pretend you have money if you get 4 sixes in a row.

Maybe it will work if the "money" you are transacting is worthless. Like worth almost one iota.

The claim "it doesnt work without centralized servers" hasnt been refuted, you have just lowered the bar significantly.

[Come-from-Beyond]

Is this a joke?

You think people will allow machines to spend/receive money and rely on only a probability that their transactions will succeed? And rely on another probability that they have the "global" tip and are not sitting in a cluster? Why not make this simpler and just roll a dice and pretend you have money if you get 4 sixes in a row.

Maybe it will work if the "money" you are transacting is worthless. Like worth almost one iota.

The claim "it doesnt work without centralized servers" hasnt been refuted, you have just lowered the bar significantly.

You were banned from IOTA thread for trolling, why do you think I'll take your words seriously here?

[iamnotback]

Iota has an entire different consensus than Byteball, and it doesn't work without centralized servers.

Do you know Java? Can you demonstrate how consensus of IOTA without centralized servers breaks once https://github.com/Come-from-Beyond/Tangle simulator is ready?

We already had that discussion. The Monte Carlo model (from the white paper) assumes that all payers and payees employ the same algorithmic strategy for disambiguating double-spends. But there is no way to enforce that assumption without centralized servers.

Rather than argue, remove the centralized servers and let's observe.

[Come-from-Beyond]

We already had that discussion. The Monte Carlo model (from the white paper) assumes that all payers and payees employ the same algorithmic strategy for disambiguating double-spends. But there is no way to enforce that assumption without centralized servers.

That link contains an example with completely random picking of tips.

[iamnotback]

We already had that discussion. The Monte Carlo model (from the white paper) assumes that all payers and payees employ the same algorithmic strategy for disambiguating double-spends. But there is no way to enforce that assumption without centralized servers.

That link contains an example with completely random picking of tips.

Random picking is not modeling an attacker who is using a strategic algorithm.

Again instead of arguing about models, remove the centralization and let's observe.

[Come-from-Beyond]

Rather than argue, remove the centralized servers and let's observe.

OK, let's postpone the final dispute to the moment when we switch Coordinator off. Obviously, we can't do it right now just to prove the point to you.