Gold collapsing. Bitcoin UP.

[Zangelbert Bingledack]

@jonny1000 It looks like we're in luck: the current form of the Classic BIP reads as follows (emphasis mine).

Activation: 75% hashpower support trigger, followed by 28-day 'grace period'

Miners express their support for this BIP by setting the fourth-highest-bit in the block's 32-bit version number (0x10000000 in hex). The first block with that bit set, a timestamp less than or equal to the expiration time, and with at least 750 out of 1000 blocks preceding it (with heights H-1000..H-1) with that bit set, shall define the beginning of a grace period. Blocks with timestamps greater than or equal to the triggering block's timestamp plus 28 days (60*60*24*28 seconds) shall have the new limits.​

 

[jonny1000]

Anyways, I drew a picture to explain one more time. The most important idea is a node will eventually accept an excessive block (i.e., a block larger than its block size limit) once enough proof-of-work exists between its chain tip and the longest potential chain tip (this picture illustrates a node with an "acceptance depth" of 3 (or 4?)

It was explained to me that that the "acceptance depth" or n was totally customizable by the user. I would set n to infinity. Your assume that people would choose 3 or 4, because that is what you want. However please make sure your proposal reflects what others will want, which is a more conservative attitude to blockszie increases than you. Some users will set n as infinity and some as 3 or 4. Therefore there is no plan or even aim of obtaining consensus between these two groups. They simply end up on different chains. Therefore BU is not part of a consensus system, it just fudges the problem by letting everyone set their own rules, such that they all operate on different chains.

Please note, even if n had a maximum value of 4 I still have many issues with it, for example helping double spends with 4 confirmations.

 

[Peter R]

"The fundamental nature of bitcoin is that it achieves decentralized consensus by longest chain" -- @Melbustus

Longest valid chain...

This is a great example of the difference in perspective between the Bitcoin Unlimited folks and the BS/Core folks.

Bitcoin Unlimited philosophy: The valid chain is the longest persistent chain (or the most-likely-to-win chain as per @Roger_Murdock ).

BS/Core philosophy: The valid chain is the longest valid chain (I think I stole this from @digitsu)

The difference is where subjectivity enters the system. With Unlimited, subjectivity enters the system in real time as every miner and node decides which blocks to extend and which to orphan.

With Core, subjectivity is in the definition of validity itself. The longest chain specifies only the order of the transactions. What makes a chain valid can not be described objectively.

 

[jonny1000]

"The first block with that bit set, a timestamp less than or equal to the expiration time, and with at least 750 out of 1000 blocks preceding it (with heights H-1000..H-1) with that bit set, shall define the beginning of a grace period. "

That is great, thankyou very much for the quote. The only way to get to 751 out of 1,000 is to cross the threshold of first getting to 750, however once 750 is reached, Classic activates. Therefore the "at least" remark is redundant and misleading, and should be withdrawn.

If you can show me how to get to 751, without first activating at 750, I will apologies and withdraw my assertion that Classic is an attack.

For example a much better idea would be to have a voting window, where at least 75% Classic support is required in that window.

Now you have a simple choice, dismiss what I am saying as me lying and trolling or listen to what I say and withdraw the Classic proposal and put forward a new proposal that can and will activate, as it gives strong consensus a chance.

[doublepost=1464056868][/doublepost]

This is a great example of the difference in perspective between the Bitcoin Unlimited folks and the BS/Core folks.

Bitcoin Unlimited: The valid chain is the persistently longest chain (or the most-likely-to-win chain as per @Roger_Murdock ).

No it is not. With BU there are still validity rules. Under BU if there is a longer chain, but with a block that has a block subsidy of 52btc, it will not be regarded as valid by BU clients.

You keep misunderstanding this as a difference in perspective. That is false, actually I think the longest chain idea would be pretty cool. It is just simply not the case and that is a fact, not a perspective.

 

[Peter R]

It was explained to me that that the "acceptance depth" or n was totally customizable by the user. I would set n to infinity. Your assume that people would choose 3 or 4, because that is what you want. However please make sure your proposal reflects what others will want, which is a more conservative attitude to blockszie increases than you. Some users will set n as infinity and some as 3 or 4. Therefore there is no plan or even aim of obtaining consensus between these two groups. They simply end up on different chains. Therefore BU is not part of a consensus system, it just fudges the problem by letting everyone set their own rules, such that they all operate on different chains.

Please note, even if n had a maximum value of 4 I still have many issues with it, for example helping double spends with 4 confirmations.

Great! It sounds like you now agree that nodes will eventually converge upon the same chain for all finite values of the acceptance depth.

By the way, everyone here already agrees that if the acceptance depth were set to infinity, then non-convergence becomes possible. We allowed this because (a) it makes it possible to emmulate Core or Classic's block size rules exactly, and (b) we believe the users are not stupid and that it is OK to allow them to adjust their advanced settings as they please. Most users simply use the default settings (which is an acceptance depth of 4).

EB = Excessive block size, AD = acceptance depth.

 

[AdrianX]

Are there any more moderate people on this forum who would like to explain why Classic locks in exactly 25% opposition at the time of activation?

Do you mind quantifying exactly what that 25% are fundamentally opposed too?
[doublepost=1464057499][/doublepost]

Again, I would really appreciate if some of you with a more reasonable attitude would speak up against this type of comment.

How about some intellectual honesty, have you ever criticized Bitcoin Unlimited until now? And addressing the facts I've presented that dispute your position.

 

[theZerg]

@jonny1000 It looks like we're in luck: the current form of the Classic BIP reads as follows (emphasis mine).

Activation: 75% hashpower support trigger, followed by 28-day 'grace period'

Miners express their support for this BIP by setting the fourth-highest-bit in the block's 32-bit version number (0x10000000 in hex). The first block with that bit set, a timestamp less than or equal to the expiration time, and with at least 750 out of 1000 blocks preceding it (with heights H-1000..H-1) with that bit set, shall define the beginning of a grace period. Blocks with timestamps greater than or equal to the triggering block's timestamp plus 28 days (60*60*24*28 seconds) shall have the new limits.​

See what I mean... He tricked you into doing his homework. You just wasted 5-15 minutes of your life .


> The only way to get to 751 out of 1,000 is to cross the threshold of first getting to 750,
> however once 750 is reached, Classic activates.

And as with setting BU to "infinite" and therefore "proving" that it does not converge he resorts to trivialities and pedantic details. The admission that will never pass his lips is the acknowledgement that this qualification is for the one idiot who deliberately tells BU to never converge by using the "infinite" setting, the rest of the network is fine.

And he completely ignores the earlier post explaining that it in fact does NOT activate at 750 but 28 days AFTER the 750 so therefore it could easily activate at 800, 900 or 1000, or even 10.

I laid his techniques out in my OP -- now you see the examples...

 

[Peter R]

No it is not. With BU there are still validity rules. Under BU if there is a longer chain, but with a block that has a block subsidy of 52btc, it will not be regarded as valid by BU clients.

You keep misunderstanding this as a difference in perspective. That is false, actually I think the longest chain idea would be pretty cool. It is just simply not the case and that is a fact, not a perspective.

This is why I said the persistently longest chain (or the most-likely-to-win chain as per @Roger_Murdock). Furthermore, I was referring to the philosophy rather than to the present implementation of BU.

Imagine that some miner mines a 52 BTC block such that it becomes the temporarily longest chain. This would still be rejected because it is not also the most-likely-to-win chain. Why should I mine on it if I'm pretty sure that it's a stupid thing to do?

Now imagine that someone mines a 2 MB block tomorrow, again making it the temporarily longest chain. Since I'm pretty sure that no one else is going to mine on top of it, once again I'm not going to bother either.

Finally, image a future where the majority of nodes are running Unlimited and, for example, bitnodes.21.co shows that 90% of nodes accept blocks up to 8 MB. Furthermore, imagine that the miners have made statements that make me confident that they will now mine on blocks up to 4 MB. Now, when that first person mines a block greater than 1MB (e.g., a 2 MB block), it makes a lot of sense for me to mine on top of it, since I now believe it is the most likely to win. Why would I mine on the previous block instead if I'm confident the majority of the network has accepted and is working on extending the new 2MB block?

 

[Epilido]

I just thought some of you wanted productive dialog with the strong economic majority. Some of you, mentioned you appreciated that. refraining from insults insures the dialog continues and is productive.

No one knows who the economic majority is. If we did all of the discussion would be over. The reason that you continue to try to convince others (and we try to convince you) that you (we) are right is because there is no clear method to determine the economic majority. A small number of people are in control of the reference client and a small number of people control the vast majority of mining. These people are clearly not the economic majority. This centralization is not allowing change to the reference client.

My concern is that by the time you, core and the controlling miners figure this out the economic majority will have already left.

Productive conversation starts with both parties not intentionally being obtuse. Ie you have repeatedly stated that classic will activate with 25% opposition. The reality is that if there was 25% explicit opposition and no single vote for classic it would not activate. So your above statement is clearly ment to be non productive.

 

[albin]

See what I mean... He tricked you into doing his homework. You just wasted 5-15 minutes of your life .

The nature of his comments suggests to me that he's playing to the lurkers.

He's trying to make it look like, when detached from the broader context (and he produces that result by fillibustering w/ arguments no casual observer would even bother reading), that he's being bullied by a bunch of "extremists" whenever he can't back up anything he's arguing.

 

[jonny1000]

And as with setting BU to "infinite" and therefore "proving" that it does not converge he resorts to trivialities and pedantic details.

It is a pedantic detail to you because you think setting n = infinity is stupid. It is not a pedantic detail to me and economic majority, who think infinity is the setting they will run.
[doublepost=1464059005][/doublepost]

The reason that you continue to try to convince others (and we try to convince you) that you (we) are right is because there is no clear method to determine the economic majority.

Please wake up to reality. The follwoing do not support Classic:

• 81% of nodes
• 95% of miners
• 85% of developers
• 85% of coin holders, in the only polls which exists

Classic loses on every metric with any basic sybil resistance. Please objectively look at these numbers. The next step is to listen to the Classic opponents and understand why they oppose it rather than making up your own ideas. The reason for the opposition is that Classic locks in 25% opposition at the time of activation, an idea which many people oppose.

 

[albin]

This demonstrates the brilliance of Classic to force the issue.

Classic is the bare minimum least controversial possible greatest compromise possible capacity increasing-hardfork, and looking at the insane grasping at straws that have to be imagined to justify opposing a blocksize bump that literally nobody credible can possibly argue is dangerous.

Toomim went to miners, asked them what would be reasonable, 2 MB was the lowest common denominator, and now look how complicated and dramatic it gets in reaction.

And to go even deeper, look how insane and convoluted this cornmaze to frivolously critique BU is getting. Basically BU is not a consensus solution because if you set it to behave like the current fixed blocksize consensus rule, it behaves like the current fixed blocksize consensus rule. I kind of admire the shamelessness of that argument.

 

[Zangelbert Bingledack]

The only way to get to 751 out of 1,000 is to cross the threshold of first getting to 750, however once 750 is reached, Classic activates. Therefore the "at least" remark is redundant and misleading, and should be withdrawn.

If you can show me how to get to 751, without first activating at 750, I will apologies and withdraw my assertion that Classic is an attack.

Naturally if the 750/1000 test is taken on a continual basis (is it?), there's no reason for the "at least" provision and it does seem like a typo.

However, your claim was that Classic could not activate unless exactly 75% of the hashing power supported it (and not 80%). Obviously we could get to 750 if we had 80% of hashing power supporting it. Sure, the BIP wouldn't have any way of knowing whether the hashing power in support was really 75% or 80%, but that's a different issue. It is nothing like "locking in" actual support at 75% (or locking in opposition at 25%), which is what you were saying.

In other words, it seems you are going back and forth between thinking of the condition as "75% of hashing power endorsing Classic" and "750/1000 blocks endorsing Classic."

For example a much better idea would be to have a voting window, where at least 75% Classic support is required in that window.

OK, but it looks like the claim you wanted to make is actually more like, "The 75% threshold only needs to be reached once, and could happen by random chance even if we only have, say, 60% support. It should be over a window where the 75% threshold is maintained continually for a certain period."

(But this seems to have the opposite problem: even if we have 90% mining support, we could fall below the 75% threshold once by random chance in the voting window and fail to activate.)

 

[Peter R]

"And as with setting BU to "infinite" and therefore "proving" that it does not converge he resorts to trivialities and pedantic details. The admission that will never pass his lips is the acknowledgement that this qualification is for the one idiot who deliberately tells BU to never converge by using the "infinite" setting, the rest of the network is fine." -- @theZerg

It is a pedantic detail to you because you think setting n = infinity is stupid. It is not a pedantic detail to me and economic majority, who think infinity is the setting they will run.

@theZerg: maybe we need a warning sign?

 

[Zangelbert Bingledack]

It is a pedantic detail to you because you think setting n = infinity is stupid. It is not a pedantic detail to me and economic majority, who think infinity is the setting they will run.

Are you saying you think that, or just that the economic majority thinks that? And why, if the economic majority is a majority, would they expect the majority of themselves(?) to do something stupid? EDIT: And also, how would you propose to prevent people from doing something stupid? Just hide the option from them?

As for counting the economic majority, I realize your point isn't that we definitely know - for binding voting purposes - what the economic majority thinks, but that it is clear that we still have not won over the people we likely need to win over. I agree we may not have.

I say "may" because there are many problems with the counting methods you mentioned, not least of which is that many people who disagree are leaving to altcoins instead of Classic. Also, there's the "waiting in the wings" phenomenon, inertia, laziness, and other things that have been mentioned many times.

@all Remember that @jonny1000 supports much larger blocks and a large part of his argumentation here is about strategy to help us get the blocksize increased. (The other part is where he actually disagrees with many of us, regarding the perceived need for "strong consensus" and doubts about BU.)

 

[Epilido]

Please wake up to reality.

Now who is being insulting?

The only way that I can think of to determine the economic majority is to have all users move all of their coins with a vote attached. This is something I am unwilling to do for personal privacy reasons. This would also likely be much more difficult to coordinate than any hardfork.

 

[Zangelbert Bingledack]

To give @jonny1000 a breather as we're bombarding him with content, does anyone know what happened to @Justus Ranvier? His last post was two weeks ago and I see he deleted his reddit and twitter accounts.

 

[bluemoon]

@Epilido

The economic majority signals all the time. At the moment it is moving to fiat, ethereum, and other altcoins. The demand for bitcoin is static or declining.

 

[albin]

@Zangelbert Bingledack

It looks like he's been doing a lot of work on github recently.

 

[Melbustus]

@Zangelbert Bingledack - I noticed that too. He got quite a bit of flak for his take-sides post on reddit and disappeared shortly thereafter. I hope he returns; he has a unique clarity of insight and writing that I've always found refreshing and valuable (whether or not I agree on the particular point being discussed).