NXT :: descendant of Bitcoin - Updated Information

[Eadeqa]

community input needed:

Unique names for assets? For example : BLABLA (only issued once)

Or non unique names for assets?

For example: BLABLA 12423434 and BLABLA 343434334  (you only choose BLABLA, the rest is the asset id - automatically generated)

(Perhaps, I'm not sure, the number can be shortened..)

If you have another idea, please specify.

I like what James proposed, no need to change Nxt core. When account creates first asset, client can create poll for the account and attach poolId end of every asset description that is created by that account. And client can get feedbacks from pollId to measure trust for that accountID and present it to user on AE window.

Also, you can restrict users from buying assets that are in not the users trust list.  That will force the users to do research and verify that they are adding a legitimate issuer to the trust list. This will cut down most of the fake/scam issuers.

[msin]

Thanks for the personal insult El3k0n.

Nxt Funds have NXT. If NXT is worth 1/32 of what it is worth today, then Nxt Funds have 1/32 of the actual dollars or Euros or gold or whatever to pay real devs, server costs, etc.

msin, I agree with you on every point you made except the part where you said you wouldn't mind if the Nxt market cap went to $1 mil.

Nxt whales have already taken $18 million of market cap out of Nxt from 50 million to 32 million today. Nxt needs new investors and needs the market cap to grow NOW. Nxt needs that market cap (price of NXT) to grow in order to attract new investors.

NXT Whales. Stop Dumping and give Nxt time to release AE. Thanks. You've taken your profit. Now invest in the future of Nxt.

Only idiots care about market cap. It's a meaningless number.

Yes, of course, I agree with spending power.  But we need to focus less on the MarketCap of Nxt at this time and more focus on advanced features, including CFB's MMORPG!

[Ola]

[Solution for Asset Naming Issue]


First off, any changes in how Asset Names are handled in the core will delay release of AE. It also impacts all the code written against the current spec. So, any solution that doesnt require NXT core changes wins any ties.

...

James

Good post, I think we need something like "verified accounts" in Twitter. The asset issuer would need to give up his anonimity. This way we know who he is and he is less likely to scam. Verified account status can be taken away too.

Very good well reasoned post james...We need to take this proposal or something similar very seriously

@wesleyh I think a verified "linkedin" Account will provide more credibility and accountability than twitter. Anybody can create a twitter account and post a bunch of tweets whereas on linkedin, your network actually kind of proves who you are and people who have worked with you in the past, because the process of connecting with other profiles requires the context of your familiarity with them and email contact. Also maybe it should be that that, verified asset issuers public information is encrypted and stored, not visible to the public unless the issuer is involved in some kind of foul play.

As for Unique names or non unique names for assets? again James is right on the money: there is no need to change Nxt core. Unique names similar to the dns model would be sufficient. What could be more valuable is the ability to destroy or sell these asset names to other accounts. What happens when company 'B' whats to sell its establishment to company 'A'? this should not be impossible to do.

[igmaca]

It should not only tied to the size of  the blockchain but also to the computational effort a node must perform to include this transaction.

EXACT

"Keep incentivized network"

interesting... what do you mean by computational effort?  It seems I am missing an aspect of something here.

must be distributed the cost of maintaining stimulated the "green, secure and decentralized network"

for example

Nxt% Fee per current transactions
Alias ​​Fee services (For example per bytes)
Nxt% Fee per AE transactions

and so on ....

How The network is worth to maintain?

How many minimum nodes are needed?
How much electricity does it cost?
How many minimum transactions contemplated

Well explained.

I dont understand a bit how these correlate to any defined computational effort.  How are some transactions more computationally intense than others?  Also, consider that AM are periodically purged, so perhaps their fee structure should be a little different somehow?  (smaller, larger, not sure)

FTR, I am 100% against charging fees based on amount of NXT being transferred.

Yes of course some services like AM it can indexed per bytes but other
can be indexed per amount transacted

all AE like  Cryptsy, BTC, Bitstamp use % amount transacted
We're going to reinvent the wheel?
why not?

what this clear that will be a lot more economical to maintain than BTC network or clones

[opticalcarrier]

I dont understand a bit how these correlate to any defined computational effort.  How are some transactions more computationally intense than others?  Also, consider that AM are periodically purged, so perhaps their fee structure should be a little different somehow?  (smaller, larger, not sure)

It is just effort to be done by a node. That effort is related to how efficient the implementation is. Each machine instruction of that node requires electricity. So, storing the blockchain, broadcasting the blockchain, creating new blocks etc. etc. everything of it requires electricity. Nothing is for free.

That is the node provider needs to be reimbursed adequately by the amount of effort he has put in to build up and maintaining his node.

One could consider the fees and the expenditures of a node provider as an implicit market. That is: expensive nodes require higher fees. People want lower fees, so only the nodes that can afford these low fees will process the incoming transactions. As node providers want to people to send their transactions to their nodes, they will workaround until they can provide such cheap nodes. On the other hand, if no node provides such low fees, people's transactions will not be distributed or included in a block, so people are willing to raise the fees.

Supply and demand will drive the fee negotiation in the long term.

FTR, I am 100% against charging fees based on amount of NXT being transferred.

Agreed.

I just dont get it.  Consider 2 people with the same identical hardware cost are block forging nodes, but one has balance of 10million and other has 1 million.  So one forges 10x as many blocks as other.  Are you saying that the smaller one should charge higher fees? This doesnt seem feasable as it will directly conflict with our desire of instant transactions.  And I still dont see how different transactions have different computations - Wont any transaction will require full knowledge of the blockchain?

You have me completely lost. Can you provide an example model, modified with your computational model?  Like if I suggested a model of a transaction of 128bytes to have a fee of .1 NXT, then if an asset creation took 256 bytes, then its fee should be .2 NXT (yes, I know AE issue probably isnt 256bytes, and its fee is currently 1000 NXT, but work with me here for an example)

[ChuckOne]

community input needed:

Unique names for assets? For example : BLABLA (only issued once)

Or non unique names for assets?

For example: BLABLA 12423434 and BLABLA 343434334  (you only choose BLABLA, the rest is the asset id - automatically generated)

(Perhaps, I'm not sure, the number can be shortened..)

If you have another idea, please specify.

I still vote for unique names.

Other solutions are not adequate IMHO.

[lumierre]

Do you guys think that 100 BTC buy wall on bter is fake? (Yes/No/Probably)

[Eadeqa]

community input needed:

Unique names for assets? For example : BLABLA (only issued once)

Or non unique names for assets?

For example: BLABLA 12423434 and BLABLA 343434334  (you only choose BLABLA, the rest is the asset id - automatically generated)

(Perhaps, I'm not sure, the number can be shortened..)

If you have another idea, please specify.

I still vote for unique names.

unique names is very very idea. If the scammer is able to issue "google" before the legitimate "google" the legitimate google will have to use "go0gle" while the scammer will be using the right name.

The unique part should be the issuer's account ID, and forcing the user to add the ID to their trust list before they can buy that asset.

That would limit the mistakes as the user will not be able to buy the asset issued by people that they have not added to the trust list.

[ShroomsKit_Disgrace]

Do you guys think that 100 BTC buy wall on bter is fake? (Yes/No/Probably)

 

[ChuckOne]

I just dont get it.  Consider 2 people with the same identical hardware cost are block forging nodes, but one has balance of 10million and other has 1 million.  So one forges 10x as many blocks as other.  Are you saying that the smaller one should charge higher fees? This doesnt seem feasable as it will directly conflict with our desire of instant transactions.  And I still dont see how different transactions have different computations - Wont any transaction will require full knowledge of the blockchain?

You have me completely lost. Can you provide an example model, modified with your computational model?  Like if I suggested a model of a transaction of 128bytes to have a fee of .1 NXT, then if an asset creation took 256 bytes, then its fee should be .2 NXT (yes, I know AE issue probably isnt 256bytes, and its fee is currently 1000 NXT, but work with me here for an example)

Alright.

Say there is a transaction A and B. A is 100 MB large and B is 10kB tiny.
Nodes can and will refuse A just because it is toooo big EXCEPT A gives the right incentive to verify, store and re-broadcast A. The fee is this incentive.

Each node can choose with pricing model it uses:
 - fees should depend only on payload
 - fees should depend on the amount of time it might require: bigger transactions like A need more time, encrypted messages might need more time, asset transactions might need more time THAN a simple transaction
 - fees could depend on both

Node providers (that is a guy running a set of nodes) can freely decide which transactions his nodes will process.

Does that make it clearer?

[ChuckOne]

I just dont get it.  Consider 2 people with the same identical hardware cost are block forging nodes, but one has balance of 10million and other has 1 million.  So one forges 10x as many blocks as other.  Are you saying that the smaller one should charge higher fees? This doesnt seem feasable as it will directly conflict with our desire of instant transactions.  And I still dont see how different transactions have different computations - Wont any transaction will require full knowledge of the blockchain?

You have me completely lost. Can you provide an example model, modified with your computational model?  Like if I suggested a model of a transaction of 128bytes to have a fee of .1 NXT, then if an asset creation took 256 bytes, then its fee should be .2 NXT (yes, I know AE issue probably isnt 256bytes, and its fee is currently 1000 NXT, but work with me here for an example)

Ah, and just for the record:

there are two types of transactions:

raw transactions -> still not included in a block
transactions in blocks -> included in a block

A node provider can decide how much fee is necessary to cover all the expenditures of the maintenance of his nodes.

[Vega]

Do you guys think that 100 BTC buy wall on bter is fake? (Yes/No/Probably)

Define fake.

[ChuckOne]

unique names is very very idea. If the scammer is able to issue "google" before the legitimate "google" the legitimate google will have to use "go0gle" while the scammer will be using the right name.

The unique part should be the issuer's account ID, and forcing the user to add the ID to their trust list before they can buy any asset.

Well, I see.

Sounds good so far. But the interface for the Average Joe.... How do you expect him to trust an account? There should be some way to verify that easily. Maybe, there could be a URL Joe could click on. That URL will provide Google's NXT account.

[lumierre]

Do you guys think that 100 BTC buy wall on bter is fake? (Yes/No/Probably)

Define fake.

Place a large buy order. Then sell to those who post above your buy order. When buy orders above are exhausted, remove your large buy order or move it down. The only risk is that somebody may actually fill your buy order.

[Eadeqa]

Sounds good so far. But the interface for the Average Joe.... How do you expect him to trust an account? There should be some way to verify that easily. Maybe, there could be a URL Joe could click on. That URL will provide Google's NXT account.

The user has to do some research. There would be warning in the client gui to make sure they are adding a trusted account ID to their trust list. The user will go to the right website (google, forums, research etc) and verify the account ID. Then the user adds that account ID to their trust list. After that, the user will be restricted to buy only assets issued from that issuer whose account ID exists in their trust list.

You can make it even better by not displaying other assets to the user in the client if they don't have the issuer ID in the trust list.  

This will cut down all the scam very fast. The scammers won't go very far (if at all) if their assets are not even displayed to the users.

Plus you avoid the problem of scammers registering all the good names.  

(1) No unique asset names
(2) Restrict users to buy/display assets whose account ID they themselves added to their trust list.

[igmaca]

Does that make it clearer?

beware

an issue is how the fee is charged and another is how the fee is partitioned

[opticalcarrier]

Yes of course some services like AM it can indexed per bytes but other
can be indexed per amount transacted

all AE like  Cryptsy, BTC, Bitstamp use % amount transacted
We're going to reinvent the wheel?
why not?

what this clear that will be a lot more economical to maintain than BTC network or clones

NXT AE, is, by definition, 'reinventing the wheel'.  You have to have a better reason than, 'thats the way its always been done'.  The other argument is basing everything on size in the blockchain.  Please argue your reason of % based with its own merits, or argue reasons against our suggestion.

Alright.

Say there is a transaction A and B. A is 100 MB large and B is 10kB tiny.
Nodes can and will refuse A just because it is toooo big EXCEPT A gives the right incentive to verify, store and re-broadcast A. The fee is this incentive.

Each node can choose with pricing model it uses:
 - fees should depend only on payload
 - fees should depend on the amount of time it might require: bigger transactions like A need more time, encrypted messages might need more time, asset transactions might need more time THAN a simple transaction
 - fees could depend on both

Node providers (that is a guy running a set of nodes) can freely decide which transactions his nodes will process.

Does that make it clearer?

You seem to believe NXT block generation is a free market commodity auction, where I can throw a transaction up for someone to give me the lowest price to include it into a block, and I then choose someone to forge my block.  What you are suggesting will break instant transactions, one of the big cornerstones of what NXT is intended to do.
what you are suggesting allows for collusion, not exactly ideal for our trustless model.

[xyzzyx]

OK a few more people chiming in... good lets get some traction.  Do we know what the smallest possible size transaction is?  Assuming that it is 128bytes like a NXT transfer is, IMO we need something even smaller than .1 NXT, I would prefer at least .01 NXT as smallest possible amount, then scaling up from there.  But that is a different conversation on the base transfer size/fee ratio is.

Currently a basic NXT payment transaction is 128 bytes: 1 byte for type tag, 1 byte for sub-type tag, 4 bytes for time stamp, 2 bytes for deadline, 32 bytes for sender's public key, 8 bytes for payee's account number, 4 bytes for the amount, 4 bytes for the fee, 8 bytes for the referenced transaction ID, and finally 64 bytes for the cryptographic signature.  All transaction types will have these base fields, then add more depending on the features.  For example, an AM transaction contains the above plus 4 bytes for AM buffer size, then the number of bytes specified by that size.

This base size of 128 bytes may change (get larger) in the future depending on how adding precision to NXT is done.  JL would have to clarify his plan on this.

[bitcoinpaul]

a. voting by account or by shares? (pandoras ballot box)

[bitcoinpaul]

It's funny the other day I was on my dedicated Nxt node forging computer (I often forget it's even on), it's a Gigabyte Brix that sits in the corner of my desk.  I decided to sync my Bitcoin wallet on the same machine.  When I opened bitcoin-qt I thought my Brix was going to explode.  CPU 100%, Fan running full blast, everything completely slow.  I quickly shutdown bitcoin-qt and let the machine relax running Nxt.  I then sent out some Nxt donations which went out instantly and sat back and drank a beer, wondering why BTC is so popular.

Please post on reddit