The probability of guessing ideal block size to within a power of 2 tends to zero as n the number of years out increases, and it tends to near zero quite quickly.
Miners and others think or perceive xt as too aggressive:https://www.reddit.com/r/rBitcoin/comments/3vadqm/charlie_lee_creator_of_litecoin_and_bobby_lee_ceo/ and we are needlessly creating this problem with code that did not solve the future scaling issue anyway.
Most likely we will need other hard forks anyway for obvious reasons like thinner blocks, and other improvements in the works. We can increase blocksize again then. Yes, it may have to be debated yet again, but automatic increases won't solve that problem either. And we may solve the issues in other ways such as thin blocks.
I think locking in future block size increases is a mistake.
--
You received this message because you are subscribed to the Google Groups "bitcoin-xt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bitcoin-xt+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
The probability of guessing within a magnitude of 2 the correct block size limit requirement (assuming we need one at all) tends to zero as the number of years out increases. There are two possibilities: If network growth exceeds the automatic increases in block size limit then the debate will be forced to continue perhaps on even less friendly terms than had we not put in automatic increases, because we will have to say they were not enough and then put in a hard fork anyway. If network growth is less than our chosen numbers it means that there may or may not be possible unnecessary vulnerabilities created: though I agree those may be exaggerated. In either case we have achieved nothing, nor have we ended the debate or the uncertainty. And note the probability of one of those cases holding tends to 1 quite quickly (within 3-5 years) AND in either of the two possible ways that we will certainly err with the automatic increases we have effectively prevented the miners from adoptiong xt. As was mentioned BTCC accounts for 14% of the hashing power, and they are not adopting xt because of the automatic increases in blocksize. Now, I cannot promise they will switch to xt if we change that but it wouldn't hurt to ask. So we see these automatic increases are achieving ONLY the continued delay in adopting bigger blocks. I personally would advise something like an increase to 4mb immediately and another doubling scheduled in one year. This is the best way to do it and it is also the way miners would like to see it done.
“We cannot predict what will happen years into the future, and we think a first step to 2 megabytes should be enough for now.” “I think it is urgent, because the transaction volume is increasing on the blockchain, and if the price rises in near future, it will push the transaction volume to be even higher. I think the max block size should be raised as soon as possible, but it should not be raised too much. Two megabytes is preferred as the first step.” Here he prefers two, but I believe anywhere 4mb initially would be acceptable for him, yes a bit of a guess based on reading his views. So we see they at least Wang and others are quite willing to consider block size increases, but they definitely absolutely do not like 8GB. The same goes for Charlie and Bobbie Lee who represent another 14%, they absolutely do not like being locked into 8GB.
Here is the quotes from Charlie and Bobby Lee:
Bobby Lee said: "My company BTCC actually came out publicly to support BIP100. The difference in the two proposals is very technical in nature. It has to do with how often the blocksize gets increased and whether human intervention is involved or whether it goes up automatically.
"I'm not too worried about which one eventually gets implemented. I have a firm belief our community will come together and implement the right solution that will us to grow and for Bitcoin to be more popular."
Charlie Lee added: "Some people do think that we need to play a lot more conservative. So increasing blocksize could potentially lead to a path of more centralisation and a lot of problems that could hurt the Bitcoin network. So it's better if you play it safe. So I am kind of in that camp. I think that if you increase the blocksize too much, too quickly, it could potentially destroy Bitcoin, with it becoming too centralised or the security suffering.”
When they hear 8GB blocks it's over, maybe we can explain it to them, but some have already taken a position that they won't likely change)
--
It appears Gavin is KEENLY aware that the miners do not like the future automatic increases.Anyway we can now state as a conclusion: The miners do not like the future automatic increases in block size. And anyone who says that future increases in block size is not an issue with the miners will have to refute Mike and Gavin's historical account of it. To Mike I would say: The miners are in China. Are you familiar with the internet there? It is exceedingly slow, unpredictable, and unreliable, when I was there I had a 3kb/s connection. Maybe it has improved, but this is not just a technological issue. Internet speed in China is way more complicated than just being a technical issue. In fact I would say this: Dependency on excessive internet speed is actually a vulnerability. Anyway, I am just trying to show you that the miners view is not necessarily totally irrational. They are the ones that would know about bandwidth in their own country. So again I say: future increases in block size limits are putting off the miners. That much is a fact as Gavin attests. So then because we get very limited value from guesses 10 years out it's best not to do that, unless we never want to see xt adopted anyway.
--
You received this message because you are subscribed to a topic in the Google Groups "bitcoin-xt" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/bitcoin-xt/b05mPLx1HWg/unsubscribe.
To unsubscribe from this group and all its topics, send an email to bitcoin-xt+...@googlegroups.com.
If they are changeable why can't we put in his preferred future blocksize limit? Could he not say: "It's changeable why don't you adopt my view?"
--
If they are changeable why can't we put in his preferred future blocksize limit? Could he not say: "It's changeable why don't you adopt my view?"Yes, but1. We can say the same thing: it's changeable, so why don't you adopt our view?2. Changing the code would mean a new BIP, new code, new testing, etc. For what? The difference between 4mb and 8mb as an upper limit is not that large.
4. The Chinese miners have made it very clear at Hong Kong that they won't consider running anything except Core. In fact their view is that they shouldn't have to think or have opinions on how Bitcoin works.Reflect for a moment upon the last point, and what that means.
Reflect for a moment upon the last point, and what that means.
I think locking in future block size increases is a mistake. I would not post here and instead just Reddit, but I feel xt may go nowhere unless the future block size increases are taken out of the code. I know this sounds radical, but part of my reasoning really shows my idea is not any more limited than the automatic increases are anyway. Let me explain. Firstly, there is a mistake in the assumption that we can predict future network growth. Let n be the number of years forward. I think future network growth is somewhere between 2^(n +1) and 2^(-n) as time goes forward. It looks like future increases of every other year leave us at 2^(n/2) as our guess. That leaves us with a max possible error of 2^(1.5n) so if we are six years forward we could be off by a factor of 2^9 a big amount. This basically leaves the odds that the automatic increases are correct or even within a power of 2 as best approximated by zero. Now, I am aware that the automatic increases may prevent a hard fork if they happen to be correct, but they won't be. I am also aware they do no harm compared to not putting them in because not putting in the automatic increases in block size is also an assumption. Except, and this is very important:Miners perceive xt as too aggressive, and mostly because of the future block size increases that will have to certainly be revised anyway.In fact more than 50% of the mining power, at least, has that viewpoint. Read more here: https://www.reddit.com/r/rBitcoin/comments/3vadqm/charlie_lee_creator_of_litecoin_and_bobby_lee_ceo/So while the guess as to the correct blocksize will surely be wrong by more than a power of 2, and hence in reality useless, it is completely preventing the adoption of xt by miners.Final point: We for a fact are going to need hard forks anyway to implement things like thinner blocks and other possible updates. Hard coding in future increases is just putting people off and is not solving anything. I propose an initial one time increase to 2mb or 4mb blocks. Four is more reasonable, but let the miners decide what they want. We need to take the initiative back from Blockstream Core and solve this. We can't try to solve it for all time in the future as that won't work anyway. In fact the automatic guess is not useful and is putting people off. Additionally, a complete solution may come from something like dynamically sized blocks. The best way to proceed is to make a bullet proof and error proof code and get it ready NOW, not several months out. We should for sure take out RBF, and implement a ONE TIME reasonable block size increase. To summarize my three main points:
The probability of guessing ideal block size to within a power of 2 tends to zero as n the number of years out increases, and it tends to near zero quite quickly.
Miners and others think or perceive xt as too aggressive:https://www.reddit.com/r/rBitcoin/comments/3vadqm/charlie_lee_creator_of_litecoin_and_bobby_lee_ceo/ and we are needlessly creating this problem with code that did not solve the future scaling issue anyway.
Most likely we will need other hard forks anyway for obvious reasons like thinner blocks, and other improvements in the works. We can increase blocksize again then. Yes, it may have to be debated yet again, but automatic increases won't solve that problem either. And we may solve the issues in other ways such as thin blocks.
--