Bitcoin XT code enabling bigger blocks

[unamis76]

https://github.com/gavinandresen/bitcoinxt/commit/821e223ccc4c8ab967399371761718f1015c766b

Finally some movement! Can't wait to see this out on Bitcoin XT's production release!!

[TierNolan]

So, 8MB and then doubling every 2 years.

I think block height rather than timestamp would be better for handling the doublings.

What is so special about the 11th January?  Wouldn't 1st January make more sense?

This is the fork rule change.

The max message size is increased to the maximum block size for 2 hours into the future.  This only applies to blocks though.  All other messages are limited to 2MB.

The rule is 750 out of 1000 and then a 2 week delay until activation but not before Jan 11th.

The version is 0x20000004, so I guess that is to allow another soft fork to be run at the same time.  It has some bitmask based stuff.

It seems complex to have to store the moment that the fork activates.  Even if it activates late, using the default start time for the doublings shouldn't be that big a deal.

[superpanos2]

It says doubling every 2 years for 2 years. so It will double 9 times.
8*2^9=2^12=4096 mb!!!!!!!
In 21 years from now bitcoin blocksize will be measured in petabytes.
GREAT! Let's hope thechnology advances quick enough, or we will see our
coins in a cryptocurrency museum.

[BitCoinDream]

... doubling every two years (so 16MB in 2018)

What does it mean ? It will keep on growing like 8, 16, 32, 64, 128, 256, 512, 1024...

That means we are going to have 1GB block limit by 2030 on XTcoin ?

Awesome

[TierNolan]

It has a max of 10 doublings, so 8GB blocks in 2036.  It depends on how fast internet bandwidth (and disk space) increases.  It is possible that we will all have fiber to the home at that point.

Second, it is possible to use a soft fork to reduce the rate at which blocks double.

[laurentmt]

What is so special about the 11th January?  Wouldn't 1st January make more sense?

A tribute to the republican marches supporting freedom of speech ?

[ajas]

What is so special about the 11th January?  Wouldn't 1st January make more sense?

On 1st January everyone is drunk.

There is nobody present to fix the problems due to the fork.

[siameze]

What is so special about the 11th January?  Wouldn't 1st January make more sense?

On 1st January everyone is drunk.

There is nobody present to fix the problems due to the fork.

That's when all the hilarity will ensue, and those of us that aren't using XT can drink more shots and enjoy the show. 

[unamis76]

Second, it is possible to use a soft fork to reduce the rate at which blocks double.

That's quite interesting

I already patched my client, xtnodes.com shows a few more nodes today.

Awesome! Every vote counts

[tspacepilot]

Wow, I new that XT was a fork that was for the block-size increase, but I had no idea it was doing a successive increase ever 2 years.  That seems wildly unnecessary to me and it makes me a lot more skeptical of XT than I thought I was.  Why on earth would we need 8GBs of transactions ever 10 minutes?  If everyone on earth was sending a bitcoin transaction every 10 minutes would we even need blocks this big for that?  And surely that's never going to happen.  Even if bitcoin was the only money on earth, not everyone would be exchanging it every 10 minutes.  Geezus, I often go for days without spending a dime, and I know I'm not the only one.

[digicoin]

> Wow, I new that XT

So are you new to Bitcoin too?

> Why on earth would we need 8GBs of transactions ever 10 minutes?

People need. Maybe you don't. You can ask the same silly question about why human population is so large? Moreover, 8GB is a limit, not effective block size. If nobody uses Bitcoin, effective bitcoin block size will be closed to 0.

[bitnanigans]

This is great. I should be spinning some XT nodes soon. I read that the blocks won't start growing larger immediately. Any idea when the larger blocks will go into effect?

[tspacepilot]

> Wow, I new that XT

So are you new to Bitcoin too?

Typos happen dude.  Maybe not to perfect people like you clearly understand everything, but for normal, fallible humans like me, typos happen.

> Why on earth would we need 8GBs of transactions ever 10 minutes?

People need. Maybe you don't. You can ask the same silly question about why human population is so large? Moreover, 8GB is a limit, not effective block size. If nobody uses Bitcoin, effective bitcoin block size will be closed to 0.

You can call the question silly or stupid, but it would have been more helpful if you'd used your omniscience to demonstrate why the qustion is silly.  My (probably wrong) estimation was that 8GB per block is waaaaay more transactions than the earth's population of humans could need.  If you could just take a little time to address that with something less terse than "people need", I think it'd be more convincing.

Also, as I'm sure someone with a god-like intelligence like you is aware, miners need fees---especially once the block-rewards are phased out.  Miners suggeerst that block-size limits create scarcity which drives up fees.

I'm sure you're completely correct  and I'm completely idiotic, but if you could just do a little more work to show me how, then my dumbass could at least get educated instead of just shamed.  Thanks in advance!

[TierNolan]

You can call the question silly or stupid, but it would have been more helpful if you'd used your omniscience to demonstrate why the qustion is silly.  My (probably wrong) estimation was that 8GB per block is waaaaay more transactions than the earth's population of humans could need.  If you could just take a little time to address that with something less terse than "people need", I think it'd be more convincing.

The plan is 8MB initially and the doubling every 2 years.  By 2036, it would hit 8GB.  With an 8 billion world, that works out as 1 transaction per person every 10 mins.

That isn't completely out of the question, especially if algorithms are making lots of trades.

It does seem a lot for the settlement/core system.

[humanitee]

> Wow, I new that XT

So are you new to Bitcoin too?

Typos happen dude.  Maybe not to perfect people like you clearly understand everything, but for normal, fallible humans like me, typos happen.

> Why on earth would we need 8GBs of transactions ever 10 minutes?

People need. Maybe you don't. You can ask the same silly question about why human population is so large? Moreover, 8GB is a limit, not effective block size. If nobody uses Bitcoin, effective bitcoin block size will be closed to 0.

You can call the question silly or stupid, but it would have been more helpful if you'd used your omniscience to demonstrate why the qustion is silly.  My (probably wrong) estimation was that 8GB per block is waaaaay more transactions than the earth's population of humans could need.  If you could just take a little time to address that with something less terse than "people need", I think it'd be more convincing.

Also, as I'm sure someone with a god-like intelligence like you is aware, miners need fees---especially once the block-rewards are phased out.  Miners suggeerst that block-size limits create scarcity which drives up fees.

I'm sure you're completely correct  and I'm completely idiotic, but if you could just do a little more work to show me how, then my dumbass could at least get educated instead of just shamed.  Thanks in advance!

If Bitcoin becomes the "internet of money" then there's going to be a lot of automated services running on top of Bitcoin.

This also leaves room for all kinds of stuff that none of us can predict. It also leaves room for a giant chunk of humanity.

Think of it this way, it's better to have way more capacity than too little. We don't want blocks completely full for very long spans of time.

[tspacepilot]

@TierNolan, I'm pretty sure I agree with you.  You seem to be echoing my skepticism about implementing something like this at this moment.  If there's really such a case as 1 transaction per person per ten minutes in 20 years, surely that's something we might want to address when, say, 1Mb blocks actually are getting full.

@humanitee I agree there's a lot that might happen that we can't predict.  But I guess I'm sorta in a state of shock to find out that XT is about pushing the block size to 8GB---wow, I thought 20MB was controversial.  I'd be for an increase to 20MB but this increase to 8GB seems a little, idunno, outlandish.

[TierNolan]

@TierNolan, I'm pretty sure I agree with you.  You seem to be echoing my skepticism about implementing something like this at this moment.  If there's really such a case as 1 transaction per person per ten minutes in 20 years, surely that's something we might want to address when, say, 1Mb blocks actually are getting full.

I think the decision should be made earlier than necessary.  The ideal would be finishing this discussion two years ago.  

They could have said that they will make the decision by the end of 2013 and have it take effect on the end of 2015 (assuming some kind of compromise).

January 2016 could still have been the switch point, but there would have been lots of time for people to upgrade.

The practical problem is that it is easy to say we should spend more time discussing things and that is what people did.  It is hard to get a decision when the problem is years away.

At this point, the decision has to be made and there is no clean way to make it.  The only option left is one that is risky all on its own (a contentious hard fork).

Increasing the size of the blocks could damage bitcoin, but not doing it could also damage things.  There is a simple disagreement about which is the greater threat to bitcoin and both sides are willing to risk the (potential) chaos of a contentious hard fork due to the risk as they see it.

@humanitee I agree there's a lot that might happen that we can't predict.  But I guess I'm sorta in a state of shock to find out that XT is about pushing the block size to 8GB---wow, I thought 20MB was controversial.  I'd be for an increase to 20MB but this increase to 8GB seems a little, idunno, outlandish.

It is an increase to 8GB over 20 years, which is different.  In 2036, bandwidth to the home could be high enough to handle it.

I think overestimating things is better than under estimating.  The block size limit is the maximum the blocks can be.  Miners can still set it lower by soft fork.

The reason to have it increase over time is so that new hard forks aren't needed every year.

[jl2012]

If there's really such a case as 1 transaction per person per ten minutes in 20 years,

8GB is about 1.5 tx per person per day, assuming 250bytes/tx and 7 billion people

[tspacepilot]

If there's really such a case as 1 transaction per person per ten minutes in 20 years,

8GB is about 1.5 tx per person per day, assuming 250bytes/tx and 7 billion people

Thanks for offering the numbers on this.  I just want to add that while we don't know what will happen with bitcoin in the next 20 years, I think everyone agrees that the idea that the world's population is going to remain flat is very unlikely.  Looks like US Census Beaureau and United Nations estimate 8.6Billion and 8.7Billion (respectively).  So, I guess that maybe it's less than 1.5 per person per day.  Is there anyone else here who thinks that the idea that every person on earth is going to send 1 Bitcoin transaction per day is a little, erm, ambitious?

I'm sure it's been talked about somewhere, but I haven't seen any details on it, but I'm curiuos why no one is pushing some kind of dynamic block size which is calculated based on the transaction volume.  We readjust difficulty to maintain a 10 minute block time.  Why shouldn't block-size be reajusted to maintain an average conf-time for a well-formed transaction with a given fee rate?

[Carlton Banks]

I'm sure it's been talked about somewhere, but I haven't seen any details on it, but I'm curiuos why no one is pushing some kind of dynamic block size which is calculated based on the transaction volume.  We readjust difficulty to maintain a 10 minute block time.  Why shouldn't block-size be reajusted to maintain an average conf-time for a well-formed transaction with a given fee rate?

You should create a fork to implement that. It's the right thing to do when there's a disagreement, isn't it?