Nxt source code flaw reports - page 11.

CIYAM

legendary

Activity: 1890

Merit: 1086

Ian Knowles - CIYAM Lead Developer

The Google Translation (am guessing) is not helping a great deal - whatever happened to all that crap we were told about computers "speaking natural languages" even by the year 2000. Grin

Edit: Indeed the "bing" translation is better - maybe Microsoft's inferior search engine has an ace up its sleeve after all?

Perhaps if the idea were to be to divide the fees between yourself and the previous forger (say 50/50) then you might get the best of both incentives (i.e. to publish as quickly as possible and to include as many txs as possible).

perl

legendary

Activity: 1918

Merit: 1190

I have edited my message ( use bing translator for help me ) .
Bing is better for the translation, it is hard to say but it's true.

Just for resume my solution:

Advantage:
1.) No free transaction possible
2.) Waiting for forge block is useless
3.) No need check if peer use more time than necessary to forge the block
4.) Nobody writes the amount it will pay (common sense)
5.) It is the individual interests of all the world to forge the block as soon as possible

Disadvantage:

Accept the idea that we are not paying for transactions validated
Payed for make network run (validate the previous blocks)
Why stay in a principle purely POW.

BloodyRookie

hero member

Activity: 687

Merit: 500

Hmm...really hard to understand you, perl.
Since you are from France, maybe german is easier for you than english?
If yes, you may write in german, I can translate

perl

legendary

Activity: 1918

Merit: 1190

Quote from: BloodyRookie on January 27, 2014, 04:56:19 AM

Getting the fees from the previous block sounds interesting, but where is the incentive to include any transactions if the generator doesn't get the fees from the included transactions?

I return your question:
Question: Actualy, What is the interest for a forge block without transaction account?
My response: Just past the following and run the network

There is also logic to pay
To include transactions in a block
and
Validate the previous block to make it not an orphan

More the block has transactions over there need not be an orphan

This would be new in the crypto currency.
But the NXT is innovative and think back the cryptomonnaie

Add an offset to compensate individuals who forge.
Does not add an interval in the validation of the transaction.
The Transaction is still in the same block, and at the same time.

BloodyRookie

hero member

Activity: 687

Merit: 500

Getting the fees from the previous block sounds interesting, but where is the incentive to include any transactions if the generator doesn't get the fees from the included transactions?

perl

legendary

Activity: 1918

Merit: 1190

Quote from: gimre on January 27, 2014, 01:36:58 AM

So if I understand you correctly, what you suggest is:

EvilBob waits with generating block n, cause he wants to maximize his gains,
In the meantime Alice produced block n

I want to prevent behavior

Quote from: gimre on January 27, 2014, 01:36:58 AM

compensation should be given to Alice?

The solution proposed

Quote from: gimre on January 27, 2014, 01:36:58 AM

That might not be that easy to do. There is (quite likely) chance that transactions in block generated by Alice,
will be different, than those, that EvilBob wanted to include.

Not likely , Is sure transaction are different
Change for switch the solution is quickly and safe .
only block use moment switch {

Block XXXXXXXXX0 , can not include transaction
Block XXXXXXXXX1 , compensation use new methode

If XXXXXXXXXX0 include fee, fee is add to account generator ( XXXXXXXX0 and XXXXXXXX1 )

}

And for get ammount fee :
Retrieve information from the previous block with same method to get signature to forge the current block

We must accept we are not paid for the transaction includes block but are in the previous block

If the previous block at zero fee. No one's interest to stay long on the current block.
Everyone wants new blocks with transactions forge

gimre

legendary

Activity: 866

Merit: 1002

Quote from: perl on January 26, 2014, 08:21:25 PM

I am ok all say.

Why the generator would not be paid with fee of the previous block ?

Generator, no gain if waiting
Generator, no fee free

We can not demand so somebody write his compendation.
Compensation must be written by somebody else

It is common sense

So if I understand you correctly, what you suggest is:

EvilBob waits with generating block n, cause he wants to maximize his gains,
In the meantime Alice produced block n

compensation should be given to Alice?

That might not be that easy to do. There is (quite likely) chance that transactions in block generated by Alice,
will be different, than those, that EvilBob wanted to include.

Let's say we have transactions: t₁, .. t_i, .. t_j, .. [t]_n.
EvilBob wanted to include t₁ ... t_j. And Alice generated t_i..t_n.

Now after that comes Charlie, who generated n+1th block. It will probably include txes missed by Alice (t₁...t_i-1).

So I don't think it would be easy to calculate HOW Alice should have been compensated.

opticalcarrier

full member

Activity: 238

Merit: 100

Quote from: perl on January 26, 2014, 08:21:25 PM

I am ok all say.

Why the generator would not be paid with fee of the previous block ?

Generator, no gain if waiting
Generator, no fee free

We can not demand so somebody write his compendation.
Compensation must be written by somebody else

It is common sense

very interesting suggestion

perl

legendary

Activity: 1918

Merit: 1190

I am ok all say.

Why the generator would not be paid with fee of the previous block ?

Generator, no gain if waiting
Generator, no fee free

We can not demand so somebody write his compendation.
Compensation must be written by somebody else

It is common sense

ricot

newbie

Activity: 56

Merit: 0

Quote from: perl on January 26, 2014, 03:19:50 PM

Is big avantage for me.
1.) 0 node can forged regular block before me.
I know all pub account and ammount ( I can compute other node forge my block )

2.) If waiting 3 more time necessary, I increase power my forge factor 3

You can't wait 3 more blocks because the blocks are chosen by cumulated difficulty. So if someone else (let's call him Bob) forges the block that you are holding off on forging, he'll have a lower cumulated difficulty than you. But since you're holding off your block, all other peers will think that Bob's block is the best and continue forging there. And as soon as a second block is forged on top of Bob's block, the cumulated difficulty of those 2 blocks will be higher than yours and no-one in the network will accept your block anymore.
So yes, you can hold off, but not for long.

Quote from: perl on January 26, 2014, 03:19:50 PM

3.) If forge one block. I can add 255 free Alias

Yes, at the moment, if you forge a block you can include transactions "for free".
I've already pointed that out. Unfortunately, there is no simple solution for that. Because if you would say that you can't include transactions from the account that is forging, you could just split up your sum into 2 accounts (and take the smallish hit with forging chances), and just commit the transactions from the account that isn't forging.

perl

legendary

Activity: 1918

Merit: 1190

Quote from: ricot on January 26, 2014, 11:56:58 AM

As I wrote in the post that I linked, waiting can loose you your block, if there are 2 other blocks generated before you release yours.
So the amount you can gain by waiting for more transactions to come in is limited to about 1-2 minutes. So yes, there is an advantage, but it's not that big.
(Btw: What's your native language?)

Is big avantage for me.
1.) 0 node can forged regular block before me.
I know all pub account and ammount ( I can compute other node forge my block )

2.) If waiting 3 more time necessary, I increase power my forge factor 3

3.) If forge one block. I can add 255 free Alias

My native language is french

bitcoinpaul

hero member

Activity: 910

Merit: 1000

https://bitcointalksearch.org/topic/m.4756305

Anon136

legendary

Activity: 1722

Merit: 1217

Quote from: klee on January 26, 2014, 08:33:25 AM

ALDRIN DONATE FOR THE PEOPLE WHO HELP HERE!

or sell it. no need for altruism here. rational self interest will get the job done nicely. it doesn't matter how he distributes his stake. all that matters is that it gets distributed.

ricot

newbie

Activity: 56

Merit: 0

Quote from: perl on January 26, 2014, 07:25:02 AM

Rest probleme, If my turn forge block .
I can waiting longtime for get maximum transaction in my block.

Waiting cannot losse my block.
I know all pubkey all peers and amount . I can compute when other node can forge block

Actualy , I have decompiler last release and not see code for punish bad node

As I wrote in the post that I linked, waiting can loose you your block, if there are 2 other blocks generated before you release yours.
So the amount you can gain by waiting for more transactions to come in is limited to about 1-2 minutes. So yes, there is an advantage, but it's not that big.
(Btw: What's your native language?)

klee

legendary

Activity: 1498

Merit: 1000

ALDRIN DONATE FOR THE PEOPLE WHO HELP HERE!

CIYAM

legendary

Activity: 1890

Merit: 1086

Ian Knowles - CIYAM Lead Developer

Quote from: perl on January 26, 2014, 07:25:02 AM

Rest probleme, If my turn forge block .
I can waiting longtime for get maximum transaction in my block.

Waiting cannot losse my block.
I know all pubkey all peers and amount . I can compute when other node can forge block

Actualy , I have decompiler last release and not see code for punish bad node

You *will* be punished when the final source code is released (the source code you have decompiled is *not* what is going to be final and even has flaws purposely injected into it).

So yes you can wait until your turn to forge to send a transaction but I really don't see that as being a flaw - rather it is an incentive to "forge".

perl

legendary

Activity: 1918

Merit: 1190

Rest probleme, If my turn forge block .
I can waiting longtime for get maximum transaction in my block.

Waiting cannot losse my block.
I know all pubkey all peers and amount . I can compute when other node can forge block

Actualy , I have decompiler last release and not see code for punish bad node

ricot

newbie

Activity: 56

Merit: 0

perl, if I understood you correctly, and I'm absolutely not sure about that, I think you mean an exploit that I was referring to a few pages ago:
https://bitcointalksearch.org/topic/m.4326007

An yes, if you are certain enough, you can avoid paying transaction fees if you only issue transactions when it's your turn to forge the next block.

gimre

legendary

Activity: 866

Merit: 1002

Quote from: perl on January 25, 2014, 12:11:02 PM

Not in place now
Is exploit is possible now
Actualy , not punish for make delay in forge block

Other exploit possible.

If forge current block.
I can add 32K transaction for create alias ( Max 255). I have return all fee when forge

You have not response for upgrade scoring for next next block

I have doubts if anyone understood, what you've wanted to say. ENOPARSE.

perl

legendary

Activity: 1918

Merit: 1190

Not in place now
Is exploit is possible now
Actualy , not punish for make delay in forge block

Other exploit possible.

If forge current block.
I can add 32K transaction for create alias ( Max 255). I have return all fee when forge

You have not response for upgrade scoring for next next block

Topic: Nxt source code flaw reports - page 11. (Read 113378 times)