NXT :: descendant of Bitcoin - Updated Information

[jl777]

When are we going to implement lower fees? NXT is a currency and not an investment. I'm an active forger and I want fees to be lowered for NXT's best interest. All we need is a fee low enough to secure the network and I believe 0.01 would do just fine and in the long term would be better for forgers even.

First we need NXT to support fractions, this is happening this month. After that it becomes possible to lower fee. I agree that the lower the fee the more the total revenue will be, even at .01

At .01 I can launch some services that otherwise is to expensive from transaction fees. This also makes trading on AE much less costly.

James

[jl777]

So, no feedback on "Create NXTcoins"  https://bitcointalk.org/index.php?topic=345619.msg5611246#msg5611246

I expected using a centralized pool approach would create a reaction like global warming.

Researching pool software and ypool's XPT (x.pushthrough) seems like at pretty good one to use. If anybody knows anything about it and wants to help, here is your chance!

James

[Come-from-Beyond]

Thanks!

Could you explain a bit how this GUID makes it immune?

James

P.S. Multisig next?

Each transaction created via sendMoney, etc. returns "guid". An exchange must store it instead of transaction id. Once a transaction is confirmed findObject will return actual transaction id (which can be changed by an adversary).

PS: Multisig will be implemented via AT.

[jl777]

Thanks!

Could you explain a bit how this GUID makes it immune?

James

P.S. Multisig next?

Each transaction created via sendMoney, etc. returns "guid". An exchange must store it instead of transaction id. Once a transaction is confirmed findObject will return actual transaction id (which can be changed by an adversary).

PS: Multisig will be implemented via AT.

So, using guid we can lookup the txid to verify that Evil Bob's txid he submits is different than the real one and reject him. There is no way for Evil Bob to change the GUID. 10 confirmations is recommended.

[ChuckOne]

I think NXT's focus should be on completing TF to increase the transaction speed, P2P AE and multiple blockchains.

Why do you think TF will increase transaction speed?

[ChuckOne]

Then this won't work.

As long as in the client, the user can choose on which blockchain to broadcast his desired transaction (different cost associated to broadcasting depending one the quality of the hardware supporting the blockchain), then high TPS could happen within the nxt network if the user decide to use "fast" premium quality network for specific important transaction.

True - you would use atomic cross-chain txs to move between NXT and NXG (something that AT will be able to provide).

Why not using assets for this?

[BrianNowhere]

THIS IS A DISASTER.

I just saw others have shared my sentiments on this, so I'm editing and paring down to this:

If anything 2101 people just got an important lesson on crypto password security. One of those lessons was just a lot more expensive than the others.

[ChuckOne]

I think BCNext talked about different fees for different transaction speeds (and/or security?). What are the implications of our two-blockchain-solution (I keep it simple for now, so only two)?

Maybe: We would have the slow one for the average user (minimal fee) and the high speed one for businesses (more fee).

Users should decide which asset (and therefore which chain) they trade (on).

After that decision is made the fees apply for that very chain and nodes forging on a particular chain decide which transaction to include.

[ChuckOne]

I think BCNext talked about different fees for different transaction speeds (and/or security?). What are the implications of our two-blockchain-solution (I keep it simple for now, so only two)?

Maybe: We would have the slow one for the average user (minimal fee) and the high speed one for businesses (more fee).

the speed of transaction processing is based on the availability of a processor, the time to get the transaction to it, the time to process and the time for the requestor to see the result.

I don't see how parallel block chains address any of these issues other than making reconciliation of accounts harder because transactions are spread across multiple chains...

EDIT: and I agree with BCNext, see earlier post, the faster you want the tx confirmed the higher the fee %age is a possible model but who pays the higher fee... the merchant wanting an instant confirmation or the buyer....?? both models exist

There are two notions of speed we should not confuse:

TPS and latency.

TPS can be increased almost to 100000000 if enough processor power and bandwidth is available.

Latency is 1 minute.

[ChuckOne]

[...] that TF is intended to favour a low number of very powerful forgers.

How do you come to that conclusion?

[funnynews]

Crazy idea, someone has already thought of it?
A coin as the NXT might have similar features with ebay?

1 - I own the account 111111
2 - I want to buy something from the seller who has the account 22222
3 - I send 100NXT, which would be trapped in blockchain at to confirm that I received my purchase.
4 - If I receive the product unlock the 100NXT account for 222.
5 - If not receive but gets stuck and only come back to me if the account venderdor 222 mark as not completed.
6 - Upon completion of the deal or not, we could both evaluate and add 1 point to the "reputation system" of accounts.
7 - Accounts with high reputation, could mediate situations where there was no agreement.

So instead of being added to ebay, etc ... that is the dream of any currency, would replace. To facilitate the exchange of NXT for other currencies without using exchange.

[ChuckOne]

How do you want to prevent people creating pools?

It seems to me that the comprehension problem I have right now is why would anybody want a PoS coin not to be secured by TF.

Until now, there is no such thing as TF => NXT has no TF.

With TF, every node can decide without any doubt what is wrong and what is right.
TF would not be gameable.
TF would not allow cheating.

[xyzzyx]

I don't think the official client should let people generate their password by default.

It should be like with Bitcoin, because people are used to that.  "NXT = Account/password" -  "BTC = PublicKey/Privatekey"


When people first open the client an account is automatically generated and people can receive coin immediatly . The password is saved in a "wallet.dat" equivalent. The user doesn't have to know it unless he search for it.

More advanced users could still create their own password and deactivate the wallet.dat creation for a real brainwallet.

But i think if we want a mass adoption, it need to be extremely easy to use.
The first step to mass adoption is adoption by the bitcoin crowd and those people are used to the wallet.dat. For those people, a wallet.dat is far more secure than a choosen password ( one guy just lose 400k NXT because he used a 8char password)



EDIT : In this thread there is a lot of very skilled people, very tech-savy, programmers, etc...So the discussion tend to revolve around those persons and their needs
Normal user/noob don't post here, so they are easily forgotten. But we shouldn't forgot that at the end of the day, they are the one that NXT need to convince. And for that, it need to be extremely easy to use !

I agree.

The NRS brainwallet interface was a cool experiment.  That experiment failed.

As you say, keep the NRS brainwallet inteface for advanced users.  Default users should get a strong NRS passphrase generated for them which they never see unless they go into advanced options, and can specify what ever simple password they want (that they can optionally change at any time) to open their client on their computer.

[ChuckOne]

One idea (which I will not take credit for, only restating in this context) was that perhaps 'effectiveBalance' (ie: forging power == 'leased + owned NXT') could be capped to prevent large pools.

Yes I think that should be one of the "settings" for a parallel chain (and maybe we ought to think about having such a limit on the main chain also).

And I still do not think what the purpose of this is except annoying stakeholders.

[Come-from-Beyond]

So, using guid we can lookup the txid to verify that Evil Bob's txid he submits is different than the real one and reject him. There is no way for Evil Bob to change the GUID. 10 confirmations is recommended.

Anyone can change transaction. All that u need is to rely on guid.

[godt]

Is it correct ?
http://87.230.14.1/nxt/nxt.cgi?action=3000&acc=10388

I think this short wallet number.

[bitcoinpaul]

How do you want to prevent people creating pools?

It seems to me that the comprehension problem I have right now is why would anybody want a PoS coin not to be secured by TF.

Until now, there is no such thing as TF => NXT has no TF.

With TF, every node can decide without any doubt what is wrong and what is right.
TF would not be gameable.
TF would not allow cheating.

You have to read the whole dialog from today to understand my post. You put my quote out of context.

1. I asked why some parallel chains would not want to choose TF
2. I asked how to prevent pools in these chains without TF

[bitcoinpaul]

Is it correct ?
http://87.230.14.1/nxt/nxt.cgi?action=3000&acc=10388

Why shouldn't it be?

But it is very unlikely that somebody has control over this account.

[ChuckOne]

Instant Transactions can't work without TF (at least part of its functions), so I'm going to add some TF features. I've already added API that lets to get list of next forgers. Next step is adding a special type of transaction that announces IPs for direct communication with the forgers (to get blocks, to send transactions).

Nice to hear.

[Come-from-Beyond]

Until now, there is no such thing as TF => NXT has no TF.

Not 100% correct. We have 50% of it.