Version 0.3.18

This is the first update to Bitcion that I'm not jumping on and grabbing right away, because I think the choice of checking transaction types in blocks is not necessarily a good thing... at least while the question is undecided in the community.  It does impact the network but in a subtle and political way.

At the very least I wish that Satoshi would at least express his view on what he thinks about putting extra data into transactions beyond pure financial data into transactions, and I think this is only going to start an arms race for those who want to play with scripts and those who are trying to keep scripts "pure".  All that has happened here is to simply lay the gauntlet down to get past the "IsStandard()" check and find transactions which can pass that test but still contain other data.

ArtForz (15+% of the network), tcatm (2Ghash/s), and doublec (pool maintainer) have stated that they will not use IsStandard. Non-standard transactions will therefore still be possible.

When I have time, I will produce a patch for Bitcoin that will remove all non-network-enforced restrictions on transactions, so that other miners can easily also be configured to include these transactions.

Please start your own block chain, rather than forking the network like this.

I hope it is clear the havoc that your push to include non-currency data is causing.

There's been exactly one block with non-standard transactions:
  http://blockexplorer.com/b/71036
... and it contained no fees.

HOWEVER:  jgarzik, you're over-reacting, too.  This will not cause a block chain split; all clients will accept blocks containing non-standard transactions as valid.  Most (many?) just won't put non-standard transactions in blocks that they generate, and won't relay them.  There will be no havoc.

No, this isn't going to cause a block chain split, it is going to cause a fork of Bitcoin.  I just hope that is thought through very well, and I don't think it has been.

Since transactions aren't being forwarded that fail the IsStandard() algorithm, those who want to put "non-standard" transactions into a bitcoin chain must either hard-connect to a miner who would accept these transactions or force some sort of ad-hoc way of setting up a second tier in the network to find nodes that will accept these kind of transactions AND likely put them into a future block.

That only one block may fail at the moment with this test is ignoring the current discussion where conceivably almost every block in the near future may have some form of these "non-standard" transactions.

Transaction fees will pay for the generation of the chain in the future, therefore if the BitDNS guys (or any other group) want to include carefully crafted transactions in the chain, then they must include the appropriate compensation for the generators, or likely their transactions will be dropped.

Even later in the future, those who include extra things without compensation in their blocks will get their blocks orphaned by the chain.

I for one will upgrade my client to v.18 until their come a time when it is profitable for me to accept non-standard transactions.

No, it'll disadvantage people who refuse to pay fees. What they're storing in their transactions is irrelevant.

Rather, it will increase everybody's fees, and incentivize creation of a currency-only chain without all the data spam.

No, as no generators will want to block fee paying (profitable) transactions.

When adding non-currency data into the block chain becomes common, then the there will become a strong economic incentive to increase the block size as necessary. Thus keeping the transaction cost down.

In the long run, extra data in the chain will _reduce_ the relative cost of a currency-only transaction, because the currency-only transaction uses much less data, compared to the data transactions, per transaction.

Sorry I have to disagree with you,  the transaction fees are based upon the 'cost' to generator to include the transaction.  The cost is directly related to the amount of data needed to be included.

In the future there will be no 'hard limits' on block size, only the risk that the block will be rejected and orphaned by the chain as it is to large.

When a generator deciding what transactions to include, if the network is already profitable with very large blocks, (eg. 20mb+ as they have much non-currency data in them), then including very small 'currency' transactions have a minute cost. For example, 20,000,000 vs 20,000,010.

If the block chain is restricted to just currency data, there is much less incentive to improve the efficiency of handling large blocks data.  Therefore the relative cost will be much greater... 500,000 vs 500,010. Therefore the transaction fees required for small amounts of data are likely to be higher on a system that only induces currency data.

Yes, that is what the source says for the moment, in the future I do not expect the generators to just arbitrarily follow the current rules, rather I expect the generators to be quite discerning in what rules they enforce or ignore.

Yes, larger block will be more expensive, (eg. bandwidth, and risk of rejection costs), but as the average amount of data in the blocks increase, the _relative_ cost per data amount will go down.

consider also that it's possible to encode data into the bitcoin blockchain already, merely by specially crafting the transaction. (See the 'first' option on http://privwiki.dreamhosters.com/wiki/Bitcoin_DNS_System_Proposal#Overcoming_potential_resistance_bitcoin_developers.2Fcommunity ).

alternatively, only timestamping and proof of expense (i.e., a regular transaction with a fee) may be used, pushing off the actual data storage off to external service (see 'third' option, ibid.)

but why not make it easier on everyone and just allow say, 64 or 128 bytes of random data in a transaction? it seems that in the net, these side-band uses of bitcoin will only serve to increase the security of the network, by increasing the mining incentive, and bringing on more generation power.

I personally agree with nanotube's point here, especially in light of the fact that arbitrary data can already be disguised in standard transactions if a person is motivated enough.  I have not looked at the source code, but from what I understand I think that a much more pleasant (politically and technically speaking) solution would be to:

by default disallow non-standard transactions that exceed 128 bytes (or whatever threshold is agreeable)?

I would like to hear why the above option was thrown out by the developers.

While I agree that bitcoin transactions themselves should be the primary use of the blockchain,  a good currency need not ostracize itself either.

...unless the use of bitcoin for non-currency purposes discourages currency use.  Some uses of the network can act as an overall disincentive against mainstream use.  If people see that miners care little for currency transactions on the bitcoin network, or all the data spam increases TX fees to annoying levels, currency users will find a new network elsewhere.  If people find out law-enforcement-objectionable data such as "kiddie-pr0n.p2p DNS data" is being managed on this network, that increases the incentive for currency users to go elsewhere.

Maybe that makes some miners happy in the short term, and you happy, but I'm here for the revolutionary new type of digital cash.

And this generalized data timestamping/notary service seems like it has the distinct probability of degrading service for digital cash, if it is even remotely successful.

Not much,    I've made about 0.5 BTC from fees.  It is just too profitable to just mine for the hell of it, I get 50 BTC each time I make a block, and (if I'm very lucky, I make 0.05 BTC from fees).

About the large chain size, this isn't a issue, in the future it will be possible to 'prune' the chain with balances. This will reduced the size by a vast amount.  When extra data is included in the chain, it is likely to be forgotten by (most) the network if it stops being important.

For example, the BitDns guys will keep a relocation of a the DNS data for ages, however the banks are likely to just keep the bitcoin currency transfer data and drop the rest.

We should not trust the "free market" blindly.

The problem is that including a lot of data in a block creates costs, not just for the miner who includes the data, but for all other miners and all other users, for all eternity.

In the current implementation, the user is passing on the costs to the miner, and the miner is passing on the costs to the rest of the network. However, the user's transaction fee only compensates the miner's costs.

As long as this tragedy of the commons situation hasn't been sorted out, the free market won't fix anything.

I propose a law:

"He who makes the chain, gets to decide what it looks like"

Well I spport Satoshi completely in this matter.

Leaving a possibility to store data in bitcoin chain is an accident waiting to happen. Just wait until somebody encodes kiddie porn into the chain - it would stay there forever. And the governments would have a perfect propaganda possibility for fighting it. "Normal" people won't use bitcoin at all if it is associated with perverts, mafia, and financial scams.

Several months ago, around the time when the 0.3.9 bugs were found, I privately told Satoshi that I thought whitelisting acceptable transaction types was a better way to go, rather than blacklisting transaction types that we find out cause problems.

The danger is similar websites that try to blacklist <script> tags in HTML entered by users to prevent cross-site-scripting hacks.  See http://ha.ckers.org/xss.html for a nice sampling of how creative hackers can be.

I haven't asked Satoshi if the recent discussion of BitDNS putting extra data in the block chain swayed his opinion or if he woke up in the middle of the night and realized that a creative use of OP_SOMETHING might lead to an exploit.  I don't think it matters; I'm still convinced that whitelisting acceptable transaction types is the right thing to do.

As for "the above option was thrown out by the developers"  -- nothing has been thrown out!  Again, I haven't talked to Satoshi, but I'm open to the idea of a third 'standard' transaction type that includes extra, arbitrary data.   Lets have that discussion, implement it on the -testnet, poke at it, try to imagine all the possible ways it can be misused, try to estimate the benefits and costs... and if there's general consensus that it is a good idea, roll it into production.