AFAICT (not authoritative, please correct me if wrong on the way it works here, my understanding is a touch shaky)

The 'mining' operation is basically a race to produce a valid block for the blockchain and have other miners accept it. The hash input data for the block is transaction data. Per-protocol rules, any miner is allowed to ignore any transaction they want to. Partly this is so that they can (if they wish) concentrate on processing transactions with a decent fee attached. This is thought to become a more significant thing as the built in reward for block discovery shrinks over time.

Therefore there is no guarantee that any transaction will be processed at all, and certainly no guarantee that it will be done in a timely fashion. Though it probably will. This is part of the wait for confirmation.

If a large miner group ignores satoshidice it doesn't mean that they disregard satoshidice transactions, it just means that they won't use them as input to the blocks they're attempting to mine. Others will still use them as input, and if the successful block comes from one of these others then the whole ecosystem counts the transaction.

However, If everyone did this, then satoshidice would indeed find it close to impossible to get transactions into the system.

--edit-- obviously this doesn't address your anonymity concerns, but others have done that.

So long as there's someone who mines coins and who accepts satoshidice transactions, they will eventually be completed. Being discriminated against slows your transactions down, it does not prevent them.

The way the system is designed to deal with discriminating miners is to allow all discrimination (a miner can choose which txns to include based on any criteria he/she wishes), but diluting their choices by having enough miners.

SatoshiDice is actually very useful for the network, because it allows us to test how it functions in a saturated environment long enough in advance that we can make changes if they are needed.

Yet this action does seem like collusion which was the oringal point. I didn't know transactions could be excluded as well. Who's to say that in the future miners could collude against certain merchants?

Oh they could if they wanted, and if they knew what the addresses in use by the merchant were. This could be made difficult by the merchant continually generating new addresses. The merchant could also add a little sweetener (transaction fee), and it feels pretty likely that a proportion of the wider BTC community would overcome their (presumably) ethical concerns and process the transactions anyway.

But yes, in principle, colluding miners could make life difficult for merchants they dislike. This is one reason why a large and divers base of miners is important to the BTC system.

There is finite (about 25) number of accounts owned by satoshidice and every one of them starts with "1dice", so there is no dificulty in identyfing satoshidice

btw: bitcoins by default allow semi-anonymous (because all transactions are public; if you want to be anonymous you need to do some tricks) transfers.

Based on my very elementary understanding of bitcoin, it seems to me it would undermine trust in the entire system, as the bitcoins lost on Satoshidice may still be spent elsewhere, and entire chains of transactions may be lost. Correct me if I'm wrong.

Seems like this fundamental issue of every node needing to replicate all transactions is turning into a real problem. I expect there will be a system of supernodes at some point, and most nodes will only store hashes. Operators of the supernodes could then become targets of regulation, undermining the decentralized nature of bitcoin and eliminating it's primary purpose.

I'd love to be wrong. Anyone?

See parallel comment for my (shaky) understanding. As an analogy.... I can't think of a good analogy!

Everyone still stores the hashes, some just don't use these transactions as hash input. May cause problems with satoshidice transaction processing and yes, could result in slow transactions and potential double spends if a large proportion of miners did this.

AFAIK Satoshidice has some clever behind-the-scenes cross validation that makes a double spend impossible even without confirmations. That's how they're able to issue payments so quickly.

Does that guard against a double spend with them or a double spend on the blockchain as a whole?

i.e. could I (theoretically, if this got a lot worse) spend BTC at Satoshidice, count on them not being validated very fast because Satoshidice transactions are ignored by a proportion of miners, and spend the same BTC somewhere else?

I'm just an interested observer at this point, I have no BTC and I'm not into online gambling.

Currently, rule-of-thumb is to wait until a transaction has been verified by 5 miners. But instead of counting, most people just wait an hour. Some merchants are more thorough than others, so if you find one that doesn't wait for a transaction to be verified properly, you could certainly defraud them within an increasingly-large time window.

However, eventually the network will include one of your transactions in the longest chain, and the other will be discarded. So the double-spend will not be permanent.

> eventually the network will include one of your transactions in the longest chain, and the other will be discarded. So the double-spend will not be permanent.

The longest chain? Why would length be the determining factor?

If someone double spends a bitcoin, it could saved by one recipient and quickly be spent many times by another. The most fair way is probably to take the first spend.

Regardless, anyone ignoring a block of a transactions seems unsustainable long term if bitcoin is going to have any credibility. Seems like those with an interest in bitcoin should avoid even considering this.

Since it's possible (and increasingly likely as more miners join) that two different miners will add different transactions to their chains, the clients are all programmed to ignore all but the longest block chain. This is important because it's harder to spoof a longer block chain than a shorter one, so longer block chains are more secure.

How is it hard to spoof a longer block chain? If you intend to do so, it's trivial. You can create an unlimited length block chain and screw the short one that is perhaps the honest one.

This is one of the more disturbing defaults I've heard of with bitcoin.

Bitcoins are not anonymous, they are pseudonymous. It works, by announcing a transaction from one of your accounts ( defined by public and private key) to another account. So if you can keep your Bitcoin address separate from yourself, then you are anonymous. But there is nothing in Bitcoins which protects the identity of your address. In fact the entire point about the Bitcoin protocol is, that there is a method to find a consensus of the contents of each individual address.

This thread provides further detail and discussion on the SatoshiDice issue:

Should the bitcoin community ban the Satoshi Dice filter patch? https://bitcointalk.org/index.php?topic=150481

especially this comment from misterbigg:

https://bitcointalk.org/index.php?topic=150481.msg1599107#ms...

Goto http://satoshidice.com/ and look. They have pre-calculated addresses that all start with '1dice' and payments to these addresses are considered a bet.

The miners are not really filtering satoshidice, they are filtering all transactions that start with '1dice' this includes them and a tiny number of false positive matches.