Hi!

Quote from: crypto_zoidberg on August 05, 2014, 06:40:04 PM

Hi amjuarez!

How's it going ?!

Glad to see you here!

Quote from: amjuarez on August 05, 2014, 05:09:12 PM

Quote from: Rias on August 05, 2014, 04:31:58 PM

I was wondering what Bytecoin developers think about Boolberry's proposals on "solving CryptoNote flaws":
https://bitcointalk.org/index.php?topic=697355.0

The solution is basically the guaranteed anonymity. Here's the original quote:

So what is the deal with Boolberry propopsal? Why wasn't the solution available in the original CryptoNote implementation? Has Boolberry actually improved the protocol?

The unique solution in BBR is to add "required mixin" flag to outputs.

First of all this idea doesn't provide guaranteed anonymity to sender....

It does, because it:

Quote

...FORCES other people to use some specific level of mixin factor in future transactions....

And this is precisely showed in our presentation.

I means that an anonymity level of one specific tx doesn't depend on "forced mixin" flags in its outputs. This flag only influences future transactions.

Quote from: crypto_zoidberg on August 05, 2014, 06:40:04 PM

Quote

Sender and the transaction he creates aren't affected directly

Surely, this transaction is a way of generating outputs with guaranteed anonimity.

Quote from: amjuarez on August 05, 2014, 05:09:12 PM

The potential problem goes from possible unspendable transactions in the blockchain: you can create an output with some very unusual amount and force addressee to spend it with big mixin factor. In case forced mixin factor is greater than number of outputs with the same amount in the blockchain this transaction will be unspendable for a long time. Addressee even will be unable to send this money back. Cryptocoin protocol SHOULD NOT allow such situations.

Dear amjuarez, you probably would be surprised, but Cryptocoin (you meant CryptoNote?) protocol already allow such situations.
I know at least two different ways to generate unspendable transactions:
1. Generate transaction to fake output keys - it would be impossible to receive by anyone - you just burned coins.
2. Generate transaction with unlock_time = 0xffffffffffffffff or other big value - transaction will be received but still won't be ever spendable.

So, you should agree that this is just a question of wallet's transaction validation, that can be easily implemented as for this case as for bbr's case that you was talked.

Unlock_time is different from the issue we are talking about because forced mixin introduces additional dependency links between transactions: spendability depends on output availability in the blockchain:

- without forced mixin txes depend on each other with strict OUTPUT->INPUT links.
- with forced mixin txes additionally depend on each other with weak OUTPUT -> OUTPUT links.

Why can't you implement spendability verification before accepting tx to mempool?

Whether it makes sense to take unspendable tx into blockchain?