Version 0.3.18

[wumpus]

Let's adhere to the KISS principle and keep it at simply a digital currency. It's already confusing enough for new users without bringing the kitchen sink in, and it could affect security. You can fork the chain if you want to store alternative information.

[davout]

as long as the fundamental properties of bitcoin remain there could be a lady gaga sextape encoded in the blockchain, i wouldn't care

[jgarzik]

as long as the fundamental properties of bitcoin remain there could be a lady gaga sextape encoded in the blockchain, i wouldn't care

The Lady Gaga sextape would impose additional costs on you and everyone else, at least if you were trying to send currency while another was pushing the sextape data into the block chain.

Then there's the new user slowdown, while each new bitcoin user downloads their copy of Lady Gaga.

[ribuck]

... it could affect security ...

Let's not spread FUD. A text payload followed by OP_DROP is absolutely secure.

It's already confusing enough for new users without bringing the kitchen sink in

A new user running the standard bitcoin App won't see anything different from what they see now.

[davout]

as long as the fundamental properties of bitcoin remain there could be a lady gaga sextape encoded in the blockchain, i wouldn't care

The Lady Gaga sextape would impose additional costs on you and everyone else, at least if you were trying to send currency while another was pushing the sextape data into the block chain.

Then there's the new user slowdown, while each new bitcoin user downloads their copy of Lady Gaga.

I hope the initial blockchain download is not going to remain as inefficient as it is today

[MoonShadow]

Bitcoin is supposed to be what users say its supposed to be.

Well, this user says that Bitcoin should remain a single tool, namely a cryptocurrency.  And if I lose the vote I'll no longer be a user, so I suppose that in the end it would all work out.

[MoonShadow]

I hope the initial blockchain download is not going to remain as inefficient as it is today

Define "inefficient" in this context.  My understanding is that the p2p network isn't inefficient for it's purpose, just not particularly well suited to bootstrapping new clients.  An alternative way of bootstrapping a new client would help to limit the new user load upon the network in the future, but that's not the problem today.  The delays in bootstrapping a new client today are largely due to the massive amount of checking that the client does while downloading the blockchain.

[davout]

I hope the initial blockchain download is not going to remain as inefficient as it is today

Define "inefficient" in this context.  My understanding is that the p2p network isn't inefficient for it's purpose, just not particularly well suited to bootstrapping new clients.  An alternative way of bootstrapping a new client would help to limit the new user load upon the network in the future, but that's not the problem today.  The delays in bootstrapping a new client today are largely due to the massive amount of checking that the client does while downloading the blockchain.

That's what would qualify as "inefficient" in my view.
Well, actually, when i say not efficient, what i mean is that i run the client on a fat linux box and this one is fine. *but* i also run a windows client at office, and each time it downloads a significant amount of blocks my hardrive sounds like it is going to burst in flames any minute.

Code:

david@bankbox:~$ du -sh .bitcoin/
124M	.bitcoin/

This should not take 10 minutes to download on a reasonably optimized P2P network, not much longer.
Bitcoin's network is probably not optimized for that, maybe we should think about using the bittorrent network to distribute fat chunks of the chain.

On windows the harddrive sound is reaaally annoying, my colleagues even get curious as to wtf I'm doing with my computer
Might be the debug.log flushes or the block chain, I honestly have no idea, but what matters is the result, isn't it ?

Also it's a very frequently asked question : "when will my faucet coins show up ?" so it could be interesting to investigate.

[jgarzik]

Define "inefficient" in this context.  My understanding is that the p2p network isn't inefficient for it's purpose, just not particularly well suited to bootstrapping new clients.  An alternative way of bootstrapping a new client would help to limit the new user load upon the network in the future, but that's not the problem today.  The delays in bootstrapping a new client today are largely due to the massive amount of checking that the client does while downloading the blockchain.

bitcoin's initial block download is far too dependent on the initial peer selection at the beginning of the download.  I've gotten unlucky several times, with a peer that was achingly slow, or a peer that simply stopped sending blocks (one peer out there seems stuck at 82000 blocks for months).  In the latter case, it took bitcoin over 30 minutes before download resumed.  Disk and CPU were completely idle during that time.

bitcoin could stand to take a page from bittorrent:  compete among multiple peers, to see which one gives you the best download speed.  If there is a network delay longer than 5 seconds, try another peer.  Accidentally downloading the same block from multiple peers, during or after a network hiccup, does not harm bitcoin.  It just wastes a bit of bandwidth.

[sturle]

... introduce junk in blocks ...

If it's useful to the sender, the recipient, and the miner, then it's not junk.

Bitcoin is supposed to be currency by definition. Anything in it that is not directly currency-related can be classified as "junk".
Perhaps separate protocols can be created for sending additional information based on private/public keys that already exist in bitcoin.

++

Those who want other services integrated with Bitcoins, just make an ad hoc protocol with it's own chain.  Let the people who store information there pay with bitcoins rewarded to those who create new blocks, or something.  Just don't push it everyone who just want to pay for stuff on the net.  It should be possible to run a Bitcoin client on a smartphone or tablet without downloading Wikileaks, porn, Gaga videos, spam, ads, warez, all of usenet, and whatnot at the same time.  Please!

[theymos]

For those who wish to remove the restriction on inclusion and relaying of non-standard transactions, here is a patch on 0.3.18 to do so:
http://pastebin.com/raw.php?i=Z0DVga74

It does these things:
- IsStandard is not checked when including transactions or relaying them.
- An older, similar restriction on transactions with an unusual size or number of scriptSig OPs is also removed.
- Non-standard transactions become ineligible for free space. You will not accept non-standard transactions if they don't include at least a fee of 0.01. You will relay such transactions, however.

Generators that apply this patch will potentially receive more fees. However, I doubt anyone will actually produce any non-standard transactions for a while.

You are still Bitcoin-compliant by applying this patch. You will not end up on a separate network (unless Bitcoin is changed to make these restrictions network-enforced).

Edit: Modified version for SVN 200:
http://pastebin.com/raw.php?i=6aQiEJDz

[jgarzik]

You are still Bitcoin-compliant by applying this patch.

This patch creates a new, non-mainline network ruleset.  The "compliant" claim is somewhat specious.

[ShadowOfHarbringer]

as long as the fundamental properties of bitcoin remain there could be a lady gaga sextape encoded in the blockchain, i wouldn't care

I would.
I want a currency, and i don't want to download somebody's lady gaga because of somebody wanted to put it in the chain.

Just think what you are saying. Is this supposed to be a currency or what ? Junkyard ?

----
BTW
I hate Lady Gaga.

[ShadowOfHarbringer]

You are still Bitcoin-compliant by applying this patch.

This patch creates a new, non-mainline network ruleset.  The "compliant" claim is somewhat specious.

Depends on what you call "mainline" and what you call "bitcoin".
If "mainline client" = what satoshi proposes, because most of users trust his vision, then mainline it is.

[jimbobway]

I also support a third transaction type for timestamp hash sized arbitrary data.  There's no point not having one since you can already do it anyway.  It would tell nodes they don't need to bother to index it.

Yes it is possible already.  But it is disappointing to encourage main bitcoin chain to be used for generic data storage.

When a large amount of users are non-currency data users, currency users will face longer delays, higher costs, and decreased ability to send free transactions.  Currency users will be discouraged by these factors from using bitcoin.

All these folks talking about the glory of the free market conveniently omit that there only one, monopolized market right now -- the mainline chain.  This implies decreased competition versus many markets (==many chains) and tragedy of the commons, as everybody shoves all their data onto the main chain.

An alternate chain that encourages data storage would better serve data users, leaving the main chain to better serve currency users.

Concerning using bitcoins as timestamps...

Right now gold can be mined and used as a currency.  You can also you gold to make jewelry, build circuit boards, etc., because it has special properties.  Therefore, gold has more than one purpose which increases its value.  As of now most people believe that bitcoin can only be used as a currency.  If people start using bitcoins to timestamp their work, then every bitcoin actually has more intrinsic value because it has more than one purpose!  A bitcoin is actually not only useful as a currency but it has the ability to timestamp.  If someone could use this special bitcoin property similarly to how scientists used the conductive properties of gold, then bitcoins could takeoff in a different tangent which could further increase its value.  Is there a breakthough application waiting to be discovered using the timestamp property out there???

[FreeMoney]

Concerning using bitcoins as timestamps...

Right now gold can be mined and used as a currency.  You can also you gold to make jewelry, build circuit boards, etc., because it has special properties.  Therefore, gold has more than one purpose which increases its value.  As of now most people believe that bitcoin can only be used as a currency.  If people start using bitcoins to timestamp their work, then every bitcoin actually has more intrinsic value because it has more than one purpose!  A bitcoin is actually not only useful as a currency but it has the ability to timestamp.  If someone could use this special bitcoin property similarly to how scientists used the conductive properties of gold, then bitcoins could takeoff in a different tangent which could further increase its value.  Is there a breakthough application waiting to be discovered using the timestamp property out there???

That makes sense Jimbob.

I don't understand well enough to know what it makes sense to pay to be put in the chain so I don't know what will be in the chain in the long run, but whatever people are willing to pay the most for is what it will be. If registering domains ends up being the best use of this system I'll be confused and disappointed though.

[casascius]

I'm relatively new and find this whole network intriguing.  I come from a software development background.  My outside opinion might not mean much, here it is for what it's worth.

I think Bitcoin should be just for currency.  The whole idea of using it for DNS or for other purposes sounds like an abuse.  If the distributed database concept brought to light by Bitcoin is viable for a DNS system, then great - may it be started as a completely independent blockchain and network.  I don't see how marrying the two systems adds any value.

If when I came across Bitcoin I had also discovered it did "DNS", I'd have been turned off by that.  There is already enough new territory for the average newcomer to conquer, let alone one with no idea how software worked.  Imagine the New York Stock Exchange also traded tampons and framing nails.  That's how the idea of Bitcoin DNS looks to me.

If I understand correctly, one of the core tenets of blocks is that they be prunable to the absolute minimum size (which I understand to be 80 bytes) to hold the "proof of work" hashes necessary to establish which chain is the longest.  The idea of using them as permanent storage for convenience betrays this.

I do think there should be a spot for arbitrary data to be sent along with transactions, but I think it should be VERY limited - and by that, I mean no more than either 32 bytes, meant as a sender's transaction reference.  32 bytes should be enough to fit a short URL, or at most, a 256-bit hash.  I know somebody gained infamy for thinking 640Kbytes should be good enough for "anybody", but this isn't the same.  I think it's fair to say that 2^256 possible references should be enough for anybody - Bitcoin should not offer payloads of any kind other than bitcoins!

Besides, we expect Bitcoin users to have Internet access after all.  Bitcoin users already have the means to communicate all the data they need, by reference.

My standard disclaimer is I'm welcome to being corrected if it appears I have misunderstood the system, as I feel I have a decent grasp but I'm still a newbie.

[MoonShadow]

I do think there should be a spot for arbitrary data to be sent along with transactions, but I think it should be VERY limited - and by that, I mean no more than either 32 bytes, meant as a sender's transaction reference.  32 bytes should be enough to fit a short URL, or at most, a 256-bit hash.  I know somebody gained infamy for thinking 640Kbytes should be good enough for "anybody", but this isn't the same.  I think it's fair to say that 2^256 possible references should be enough for anybody - Bitcoin should not offer payloads of any kind other than bitcoins!

I agree, as do many others.  There is great risk in trying to adapt Bitcoin into some kind of multi-tool.  Simple tools that can be used in conjunction is what made Unix great, and both Linux & the Internet great by proxy.  Bitcoin should be a currency system, and leave the othe innovations to other tools.  The argument that some have made that another blockchain would take generation from Bitcoin (or vise versa) is a false one.  People will contribute to what is important to them.  Neither distributed.net, nor any other of the prior distributed work projects, makes any attempt to compensate participants.  If a p2p DNS is important, people will volunteer both clocktime and bitcoin donations to support it.

[ShadowOfHarbringer]

I'm relatively new and find this whole network intriguing.  I come from a software development background.  My outside opinion might not mean much, here it is for what it's worth.

I think Bitcoin should be just for currency.  The whole idea of using it for DNS or for other purposes sounds like an abuse.  If the distributed database concept brought to light by Bitcoin is viable for a DNS system, then great - may it be started as a completely independent blockchain and network.  I don't see how marrying the two systems adds any value.

If when I came across Bitcoin I had also discovered it did "DNS", I'd have been turned off by that.  There is already enough new territory for the average newcomer to conquer, let alone one with no idea how software worked.  Imagine the New York Stock Exchange also traded tampons and framing nails.  That's how the idea of Bitcoin DNS looks to me.

If I understand correctly, one of the core tenets of blocks is that they be prunable to the absolute minimum size (which I understand to be 80 bytes) to hold the "proof of work" hashes necessary to establish which chain is the longest.  The idea of using them as permanent storage for convenience betrays this.

I do think there should be a spot for arbitrary data to be sent along with transactions, but I think it should be VERY limited - and by that, I mean no more than either 32 bytes, meant as a sender's transaction reference.  32 bytes should be enough to fit a short URL, or at most, a 256-bit hash.  I know somebody gained infamy for thinking 640Kbytes should be good enough for "anybody", but this isn't the same.  I think it's fair to say that 2^256 possible references should be enough for anybody - Bitcoin should not offer payloads of any kind other than bitcoins!

Besides, we expect Bitcoin users to have Internet access after all.  Bitcoin users already have the means to communicate all the data they need, by reference.

My standard disclaimer is I'm welcome to being corrected if it appears I have misunderstood the system, as I feel I have a decent grasp but I'm still a newbie.

+ 1, you have my full support.

[The Madhatter]

I think Bitcoin should be just for currency.  The whole idea of using it for DNS or for other purposes sounds like an abuse.

*chomp*

I agree. I recently upgraded the 30 nodes I control to 0.3.18 with verbatim IsStandard() code, just to prove my support.