Bitcoin Forum

Thanks to DataVeteran's twitter page (https://twitter.com/Datavetaren), I've stumbled on the following link:

https://data.mtgox.com/api/0/bitcoin_tx.php

It shows a bunch of mtgox blockchain submissions. They all include the new transaction ID constructed by Sipa and Krapeles as disussed extensively on the GitHub thread (https://github.com/bitcoin/bitcoin/pull/3656).

Copy/Pasting the NTXID field in blockchain.info gives a series of transactions happening over the course of the past couple days. If I understand correctly this is the testing they announced they would be doing.

Not all NTXID seem to work. This might imply that not all test transactions are going through properly and, thus, the system isn't fully ready for primetime action. If this were the case, the freeze on all withdrawals could be deemed appropriate.

This is good news for people that are hoping that the exchange doesn't die and they can recover their funds.

However, it still makes me wonder: if Mtgox is solvent and the delay are just really caused because of technical issues or incompetence, why would Karpeles deny to comment about Mtgox's solvency? Why wouldn't he want to avoid paranoia and make their customers have peace of mind? The only reason I find is to actually crash the price on purpose.

I don't know much about programming, but what does it mean by the fact that all the script signatures are redacted?

Maybe he is a sociopath?

At this point, people won't believe anything they say. They may figure they might as well make millions from all the trading.

I guess all the FUDsters will say this ntxid testing is just a smoke screen.

Exactly, as would be the reaction from them saying everything was fine.

I'm not quite sure what you're asking for, so sorry if I my answer is not what you expected.

The ntxid is calculated the same way that the txid is calculated, but by removing all scripts. Why? The idea is that the scripts themselves are not signed (and cannot easily be), whereas other part of the transactions are signed. Now an attacker may make a subtle modification to the script; and since the scripts are not signed, even if the attacter does not have the private key, the transaction remains valid.

There is now two similar but different transactions. They both have the same effect: "transferring n bitcoins form X to Y" (this data is signed). But because they have the same effect, only one of them can confirm. We say that that one of them is mutated (but to a miner, who don't know the sender, he cannot tell which is which, both being equally valid). Their txid is different, because the transactions are slightly different.

If the "mutated" one confirms, it may confuse the sender if the sender only checks if a given txid is confirmed.

The ntxid is a solution to that: By blanking the scripts, the ntxid is the same for ALL transactions that have the same effect. Since this data is controlled/signed by the sender, he can now be sure that if he checks for confirmation of a given ntxid, he will know, even if it is a mutated version that confirms.

(The original txid remains useful because the nodes need a way to discern the different-but-equally-valid mutated versions one from the other.)

I would say the fact that it is still not working is just a sign of ongoing incompetence. Seems they will be in legal trouble soon enough though.

If this is serious, the withdrawals should take months to return to normal anyway, they already announced "new limits" due to "market conditions"(their own condition emminent bank run), the closed box will continue with low value goxcoins.