NXT :: descendant of Bitcoin - Updated Information

[Come-from-Beyond]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Some ppl confuse green and red...

[1714843824]

1714843824

[1714843824]

1714843824

[1714843824]

1714843824

[nexern]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Thx. But what about after clicking "Send"? As I recall wrong sends happened after correct address was set.

you mean some kind of checksum?
as soon there is a difinition for this i am adding it for sure.
the above mentioned solution is of course very simple but effective as well.
within an option tab you can set the warning level. of course the user will get
an extra confirmation dialog for yellow and red after pressing send.

a realtime verification during typing depends on the gui-latency but as soon the
account number field lost focus the number is checked and gives a visual feedback.

[jl777]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Some ppl confuse green and red...

Yes, not insignificant percentage of people are color blind. We need a specific dialog box in the red case, defaulted to NOT send.

James

P.S. What version should peercover run? 4.8 or 4.9, if 4.9 where is it?

[nexern]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Some ppl confuse green and red...

the user get's both, visual feedback and form warning.  

without an additional alert dialog, the color wouldn't make much sense of course...

[jl777]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Thx. But what about after clicking "Send"? As I recall wrong sends happened after correct address was set.

you mean some kind of checksum?
as soon there is a difinition for this i am adding it for sure.
the above mentioned solution is of course very simple but effective as well.
within an option tab you can set the warning level. of course the user will get
an extra confirmation dialog for yellow and red after pressing send.

a realtime verification during typing depends on the gui-latency but as soon the
account number field lost focus the number is checked and gives a visual feedback.

intel, the person with handle not company, had a theory that memory corruption in the JAVA could cause the destination acct to get corrupted. Another theory is that it could be a race condition of some sort that exposes an uninitialized variable, basically stuff that is outside the realm of what the client can do.

WE need to refuse to create darkNXT in the client and independently refuse to create it in the JAVA and ideally refuse to create it in the protocol. What purpose does darkNXT serve other than to piss people off because they lost small to large amounts of money. In alpha test, ok. In beta test, maybe ok but not really. After formal launch we can never create darkNXT, unless we are willing to reimburse, which does not seem likely.

Ripple does this better than NXT. Can we really allow that to stand?

Nexern, there is only so much you can do in the client.

James

P.S. Sorry CfB, we might even need a "allow darkNXT creation" API call. That is how difficult it needs to be. No laughing matter extinguishing peoples money.

[markn]

Questions for DGEX operator

2. Is the withdrawal fee for new deposits only, or does it include existing deposits? If it covers existing deposits that is quite frankly taking the piss. You can't take people's money and then say ok I'm now charging you to get it back. At least give people notice to withdraw their funds out first - I certainly saw no notice on the site.

I buy everyday small amount of NXT and from my experience I can say that since the end of December I have to pay for withdrawals of NXT.
I think that now NXT come "faster" to my NXT account (relatively, in the beginning it was sometimes 4 days now I have them till next day )

Of course at first, (when I started 24.12. FYI - I came quite late into NXT) I also tried several times to withdraw BTC, and I paid 3.2% but I wanted to test if Dgex is not honeypot where money flows in but not out.
I understand we should pay some fee, that is OK, but I would prefer to pay for every trade - it has some logic - the more you use the service the more you pay
(and btw for me it is also cheaper when I do trade manually small amounts /not automatically through API and some programme/)

[NxtChg]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Why invent crazy schemes when there is a simple, elegant solution, which covers both client AND server - checksums. The way it's implemented in bitcoin.

I don't buy the whole "oh my god performance" line of reasoning. You wrote it in Java, for god's sake

[salsacz]

CfB + jean-luc + nexern,

There are millions of darkNXT and more being created all the time. darkNXT is NXT sent to an address that doesn't belong to anybody. I am assuming this is something we don't want the end user to be able to do easily. Anyone with direct API access could do whatever they wanted, but end users should NOT be allowed to create darkNXT, especially when all it takes is a typo or memory corruption.

The solution to this is so simple I am at a loss as to why it has not been implemented. NXT sent to an acct without key will create darkNXT, there is a call to detect if an acct has a key. Why has this simple guard not been implemented?

This is a very serious flaw and even ripple has a solution to this by not allowing money to an unfunded acct. We can't let horrible flawed useless ripple be better than NXT in this area can we?

James

P.S. For those not in the real world, EVERYTIME and end user creates darkNXT, it will create at least a support incident at worst lawsuits against a vendor supporting NXT. Need I remind people that we need widespread NXT adoption.

I hope nexern will fix this issue on client side.

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

what about double sending? Simple solution is to send 1 Nxt first, then users can check the trade on the blockchain - and then make a real trade to the same account

[bitcool]

did someone finished reading the source code ......?
conclusion?

One file, all static.

Very interesting way  to use Java.

[nexern]

already wrote about this two weeks ago.
the client wil handle it this way:

1. check if address is within own address book  -> green
2. if account exist within blockchain -> yellow
3. don't exist within blockchain -> red

simple visual control.

Why invent crazy schemes when there is a simple, elegant solution, which covers both client AND server - checksums. The way it's implemented in bitcoin.

I don't buy the whole "oh my god performance" line of reasoning. You wrote it in Java, for god's sake

as soon the extra bit's are available for this, i can add it.

just to explain why a colored feedback is not only eyecandy but usefull, consider this.
the first step is to enter the address where to send the funds. when done the address
field is loosing focus because you do the next step, enter the amount. at this point you
get the visual feedback already about the account number. no need to fill out all other
fields and press send to know something is wrong with your address.
just a little thing but userfriendly and i can add it right now.

as said, just an example...

[marcus03]

Since I need to take a day off of programming, here's the first version of my, yet unnamed, NXT client for Windows.

As it is, it's really more of a block chain explorer (you can't send NXT, register aliases, etc. and you are currently not asked for secret phrases for accounts).

Unzip the archive to a directory and start the NXTclient.exe. Make sure that NRS is running locally, or enter an URL like "http://192.168.178.12:7874" when asked (it defaults to "http://127.0.0.1:7874").

The data subdirectory holds the blockchain data in proprietary format as of an hour ago. To import new blocks/transactions from the NRS blockchain into the client, click on the little Refresh icon and relax.

All the data in the blockchain explorer part can't be edited (since it's imported from and in synx with NRS), except for the column "Description" in the list of accounts.

The interface is still very, let's say functional - no bells and whistles yet.



NXTclient-v1.0.zip (28.2 MB) - Download: https://mega.co.nz/#!ONd3XCLQ!O_xx-C6os70pe41Tc3j6CmoUTEl6922SDytC8WD8lko

SHA256 checksum for NXTclient-v1.0.zip: 485735CCB66323E6F0FBF016A4D15134263D6970A9AA81E82CFF5D1AF17C9FEC


Next on my TODO list:

• Add aliases to blockchain explorer
• Send NXT function and personal account management

Hope it's already helpful for anyone (e.g. when the online NXT blockchain explorer is down).

Of course, I would not feel like a real NXTer without asking for tips...  NXT: 1758531264253431177

[Come-from-Beyond]

did someone finished reading the source code ......?
conclusion?

One file, all static.

Very interesting way  to use Java.

U haven't seen 0.2.*

[tk808]

Since I need to take a day off of programming, here's the first version of my, yet unnamed, NXT client for Windows.

As it is, it's really more of a block chain explorer (you can't send NXT, register aliases, etc. and you are currently not asked for secret phrases for accounts).

Unzip the archive to a directory and start the NXTclient.exe. Make sure that NRS is running locally, or enter an URL like "http://192.168.178.12:7874" when asked (it defaults to "http://127.0.0.1:7874").

The data subdirectory holds the blockchain data in proprietary format as of an hour ago. To import new blocks/transactions from the NRS blockchain into the client, click on the little Refresh icon and relax.

All the data in the blockchain explorer part can't be edited (since it's imported from and in synx with NRS), except for the column "Description" in the list of accounts.

The interface is still very, let's say functional - no bells and whistles yet.



NXTclient-v1.0.zip (28.2 MB) - Download: https://mega.co.nz/#!ONd3XCLQ!O_xx-C6os70pe41Tc3j6CmoUTEl6922SDytC8WD8lko

SHA256 checksum for NXTclient-v1.0.zip: 485735CCB66323E6F0FBF016A4D15134263D6970A9AA81E82CFF5D1AF17C9FEC


Next on my TODO list:

• Add aliases to blockchain explorer
• Send NXT function and personal account management

Hope it's already helpful for anyone (e.g. when the online NXT blockchain explorer is down).

Of course, I would not feel like a real NXTer without asking for tips...  NXT: 1758531264253431177

Great work! Looks great.

[NxtChg]

just to explain why a colored feedback is not only eyecandy but usefull, consider this.

Ok, if you can make it purely on the client by checking local blockchain fast and cheap - that's great and should be added.

I am just saying that this doesn't solve anything. You don't protect against typos, because there are plenty of legit cases where address will not be in the blockchain.

Then, you don't protect against unhealthy server at all, because, well, you're a client.

Checksums are used everywhere. Here we are talking about big money and the cost of failure is irreversible loss of it.

Honestly, it's beyond me why are we even arguing this.

[salsacz]

did someone finished reading the source code ......?
conclusion?

One file, all static.

Very interesting way  to use Java.

So is Satoshi author or not?

[markn]

So is Satoshi author or not?

I belive not. Satoshi is excellent in C++ not in Java

[landomata]

yet unnamed, NXT client for Windows.

May I suggest a name.... Nxt Jupiter Client ...or Jupiter for short.

[nexern]

just to explain why a colored feedback is not only eyecandy but usefull, consider this.

Ok, if you can make it purely on the client by checking local blockchain fast and cheap - that's great and should be added.

I am just saying that this doesn't solve anything. You don't protect against typos, because there are plenty of legit cases where address will not be in the blockchain.

Then, you don't protect against unhealthy server at all, because, well, you're a client.

Checksums are used everywhere. Here we are talking about big money and the cost of failure is irreversible loss of it.

Honestly, it's beyond me why are we even arguing this.

you are misinterpreting my statement here. i am not against checksum at all.
as said, this is an simple, quick and effective improvement compared to the
current client, not more, not less.

@marcus03, great work marcus. i have some question about how you have managed orphaned blocks and checkpoints etc.
will contact you via pm.

[gbeirn]

yet unnamed, NXT client for Windows.

May I suggest a name.... Nxt Jupiter Client ...or Jupiter for short.

What about Last (LST).

[wesleyh]

Hi Jean-Luc, CFB and others interested;

I'm ready with that client side implementation of an SHA256 test. I would like some people to test this. I have tested it on the latest Safari, Firefox, and chrome.

As far as I know this will only run in IE10+, earlier versions of IE are not supported.

You can download the files here: http://nxtra.org/update/update.zip

Unzip, then go to your nxt folder and navigate to webapps > root. Paste the two files there.

Now you should be able to open http://localhost:7874/update.html in the browser.

For this test you will always receive the update notice. An older version is hardcoded in. This is to allow for testing of the sha256 hashing function.

Just follow the instructions; download the client, and then drop it over the dropzone to see if it compares to the official release.

If all goes well you should see this screen:



Now also test it with another file, just a simple txt file or another zip, it doesn't matter. Just pretend this is a fake release.

If that goes well, you should see the following:



I hope this will increase the security of everyone, if you have any feedback, please share.

My Nxt ID for donations is: 3209075099254042753 - any donation is appreciated.

Thank you,
Wesley