4 ms·

Their testnet didn't fork properly either, but for some other reason I believe. It was thought to be because they had more mining power than anticipated in thei

Their testnet didn't fork properly either, but for some other reason I believe. It was thought to be because they had more mining power than anticipated in their testnet, which was discussed on the mailing list as an attack on testing. One reason why testing wasn't thorough enough was that it was under specified. There was no public discussion on how this was supposed to work. The specification was whatever the maintainer decided to merge, which changed several times during the software's lifetime. Most further discussion or counter proposals, including perhaps the most controversial one which was replay protection, were met with "that wasn't what the signees agreed to" and that anything going beyond that was off the table. But what the signees agreed to was always very unclear, as the agreement only described that a "fork" was to be activated and that something was supposed to be 2 MB in this fork. What was supposed to be 2 MB was not specified, let alone how deployment was supposed to work, and neither the protocol for signalling, fork activation and block weight. And the latter things are probably what reviewers would have concerned themselves with had this ever been a real proposal.

stale2002 9y ago
What "2MB was supposed to be" was absolutely obvious to everyone who actually signed the document. It was supposed to be twice as much as segwitz done via a base blocksize hardfork (thus, the name segwit2X). The only people who were "confused" about what it was were the smaller block trolls who didn't even sign or support the document in the first place, therefore their opinion doesn't matter. The only opinions that mattered were of those who actually supported the NYA and segwit2X. Some other details, such as replay protection, we're obvious sabotage efforts that were also proposed by the small blocker trolls. None of the NYA people were arguing for that, and once again, they were the only opinions that mattered. The other details, such as deployment date, signaling method, ect were not specified, though, and I agree that this came back to bite everyone in the ass, as proven by the fact that it failed.
- xorcist 9y ago
  No, far from it, as evidenced by the opinions of the undersigners on where to take the project from the start. And while labelling dissenting opinions trolls may make for convenient discourse, it does not help with understanding the underlying engineering. Note that the most common reason among the former undersigners for withdrawing support was the lack of replay protection (which only underscores the lack of consensus around some fundamental design decisions among the signers). To anyone familiar with post segregated witness Bitcoin an increase in base block size could be done in a number of ways. The weights for data structures could be changed, but so could the weight limit. What would be reasonable if backwards compatibility is no concern is use the same multiplier for all types of signatures, thereby removing the so called segwit discount, but this turned out not to be the design chosen. Like most people interested in Bitcoin, I only followed the NYA at a distance and have no more information than anyone else in this space, but I do know some people working with a company that took part of the agreement. I do not think there was some secret cabal behind the agreement, which some people seems to believe, but I also think that there were conflicting goals involved.

stale2002 9y ago

What "2MB was supposed to be" was absolutely obvious to everyone who actually signed the document. It was supposed to be twice as much as segwitz done via a base blocksize hardfork (thus, the name segwit2X). The only people who were "confused" about what it was were the smaller block trolls who didn't even sign or support the document in the first place, therefore their opinion doesn't matter. The only opinions that mattered were of those who actually supported the NYA and segwit2X. Some other details, such as replay protection, we're obvious sabotage efforts that were also proposed by the small blocker trolls. None of the NYA people were arguing for that, and once again, they were the only opinions that mattered. The other details, such as deployment date, signaling method, ect were not specified, though, and I agree that this came back to bite everyone in the ass, as proven by the fact that it failed.

xorcist 9y ago
No, far from it, as evidenced by the opinions of the undersigners on where to take the project from the start. And while labelling dissenting opinions trolls may make for convenient discourse, it does not help with understanding the underlying engineering. Note that the most common reason among the former undersigners for withdrawing support was the lack of replay protection (which only underscores the lack of consensus around some fundamental design decisions among the signers). To anyone familiar with post segregated witness Bitcoin an increase in base block size could be done in a number of ways. The weights for data structures could be changed, but so could the weight limit. What would be reasonable if backwards compatibility is no concern is use the same multiplier for all types of signatures, thereby removing the so called segwit discount, but this turned out not to be the design chosen. Like most people interested in Bitcoin, I only followed the NYA at a distance and have no more information than anyone else in this space, but I do know some people working with a company that took part of the agreement. I do not think there was some secret cabal behind the agreement, which some people seems to believe, but I also think that there were conflicting goals involved.