cypherdoc (OP)
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
June 04, 2015, 05:19:44 PM |
|
On another subject, the anti-expansionists think that we could hard fork in an emergency in a month. Maybe we should challenge them to prove it. Let's hard fork to 2MB (the smallest reasonable increase) or 2MB + 10% a year if they'll agree within a month. If that works, maybe I'll believe that we can do it quickly when it counts
That would have the advantage of quieting the extremists, like Mircea Popescu and his following, who refuse any change at all. Once we have made an upward change, that horse has left the stable. Plus, if tx volume doesn't simply increase to fill the space but rather miners self-limit, it will suggest that a 20MB (or 8MB) limit isn't going to mess anything up either. It will also be progressively harder to argue for the "well-connected miners torment poorly connected miners with big blocks" attack. In many ways this will be the camel's nose under the tent, as well as serving the purpose you mentioned, that Greg and Luke claim we can hotfix on the fly (if we can, great, let's do that, but if not...). it's SO obvious you cannot fork at the last minute. it's simply b/c a huge portion of the network won't get the memo to upgrade immediately. merchants will keep processing tx's that get forked off the network. chaos and confusion will occur. this is why Satoshi talked about a slow progressive "versioning" update with the final features only being enabled after 8-12 mo or so when it's clear thru monitoring that most everyone has updated. this is what we're doing right now with XT. i think it's really disingenuous for Greg to say so. yet another black mark in my checkbook for him.
|
|
|
|
Adrian-x
Legendary
Offline
Activity: 1372
Merit: 1000
|
|
June 04, 2015, 05:19:55 PM |
|
The point is that coins aren't really done and on auto-pilot. They require ongoing upkeep from lead devs.
This is a good point, and part of why I consider all current crypto coins to be not ready for prime time. When something is truly on permanent auto-pilot then we can accept it is a working decentralized system. MP's point about Bitcoin is, I think, that it should simply never be hard forked. If it fails, it fails, and perhaps is replaced by something better. But the idea of any developers having that kind of power is a fundamental failure of the concept. It's worth considering. i doubt that any cryptocoin can ever be on auto pilot as the crypto evolves as computerization advances. what is secure today won't be secure tomorrow thus requiring continual updating. the term hard fork is a bad one, imo. even after 6y, definitions amongst early adopters varies. in my mind, i think of them as necessary upgrades. they are in fact necessary over time as situations change and crypto cracking techniques mature. or even as the economic conditions change, like i think we are seeing now with the restrictions 1MB is causing. the increasing block limit movement is Gavin responding to continued lobbying by the economic majority of Bitcoin users who are acting out of conditions in the real business world. the crypto-anarchists hate this. i get their point but as i've already said, if one's fundamental unit is the full node and not the user, i think you're doing it wrong. network work squaring effects will correlate with the user, not full nodes. we see this in all comparable models; Uber, AirBnB, Facebook, Twitter, etc. Full nodes are analogous to ACH or Swift which is simply the plumbing or transmission services for users. small blocks are the ultimate in centralization. all you have to do is look at the system as it is today as a result of 1MB blocks; confined mainly to the 2 most regulated geographic regions of the world, the US & Europe. with usage still primarily by geeks. that's a recipe for heavy intervention by regualtion. who honestly thinks that Nasdaq will expand their trading systems while constrained to 3 tps? i think the Visa's and MC's are laughing at us while some of us fight hard to keep us constrained. What will happen if companies like 21 inc litter the globe with their tech built into hardware devices? How could a hard fork ever work then? Miners do what the nodes allow them to do. Nodes are maintained by network users with an invested interest in preserving the value stored in the network. The Network of users grows because Bitcoin does what it says on the can. The many supporting Gavin are not the majority, the majority in Bitcoin is the economic majority and should be typical distributed over a typical bell curve. The voters are the nodes, some developers want to keep centralized control over the majority of the nodes, some developers realize this is bad, - maintaining control its either explicit, subconscious, or subverted. KNC just launched a chip that is 800% more efficient than there previous most efficient chip. Before 21 becomes a problem, KNC are going to become the new BTCGuild or GHash.io. they will also be eclipsed when inlet, AMD or someone Taiwanese breakaway chip manufacturer thinks there is an opportunity in Bitcoin. So with the above in mind, i think that's a good idea so long as 21 are not cooperating with the Core developers who what to maintain centralized control, and if they are the problem is not 21, its centralized control of the code that runs the nodes.
|
Thank me in Bits 12MwnzxtprG2mHm3rKdgi7NmJKCypsMMQw
|
|
|
cypherdoc (OP)
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
June 04, 2015, 05:21:21 PM |
|
major dump going on. $DJI down -200
|
|
|
|
cypherdoc (OP)
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
June 04, 2015, 05:35:55 PM |
|
On another subject, the anti-expansionists think that we could hard fork in an emergency in a month. Maybe we should challenge them to prove it. Let's hard fork to 2MB (the smallest reasonable increase) or 2MB + 10% a year if they'll agree within a month. If that works, maybe I'll believe that we can do it quickly when it counts
That would have the advantage of quieting the extremists, like Mircea Popescu and his following, who refuse any change at all. Once we have made an upward change, that horse has left the stable. Plus, if tx volume doesn't simply increase to fill the space but rather miners self-limit, it will suggest that a 20MB (or 8MB) limit isn't going to mess anything up either. It will also be progressively harder to argue for the "well-connected miners torment poorly connected miners with big blocks" attack. In many ways this will be the camel's nose under the tent, as well as serving the purpose you mentioned, that Greg and Luke claim we can hotfix on the fly (if we can, great, let's do that, but if not...). it's SO obvious you cannot fork at the last minute. it's simply b/c a huge portion of the network won't get the memo to upgrade immediately. merchants will keep processing tx's that get forked off the network. chaos and confusion will occur. this is why Satoshi talked about a slow progressive "versioning" update with the final features only being enabled after 8-12 mo or so when it's clear thru monitoring that most everyone has updated. this is what we're doing right now with XT. i think it's really disingenuous for Greg to say so. yet another black mark in my checkbook for him. even if the network is lucky enough to get to the "last minute" before losing all it's customers from unconf tx's, tell me how you know you're there? apparently it will be decided by core devs like Greg. the obvious question is then, "where is the last minute?" if anything the last few weeks of contentious debate has demonstrated is that we can't decide on what the last minute will be even if we agree to wait until the last minute.
|
|
|
|
thezerg
Legendary
Offline
Activity: 1246
Merit: 1010
|
|
June 04, 2015, 05:42:42 PM |
|
On another subject, the anti-expansionists think that we could hard fork in an emergency in a month. Maybe we should challenge them to prove it. Let's hard fork to 2MB (the smallest reasonable increase) or 2MB + 10% a year if they'll agree within a month. If that works, maybe I'll believe that we can do it quickly when it counts
That would have the advantage of quieting the extremists, like Mircea Popescu and his following, who refuse any change at all. Once we have made an upward change, that horse has left the stable. Plus, if tx volume doesn't simply increase to fill the space but rather miners self-limit, it will suggest that a 20MB (or 8MB) limit isn't going to mess anything up either. It will also be progressively harder to argue for the "well-connected miners torment poorly connected miners with big blocks" attack. In many ways this will be the camel's nose under the tent, as well as serving the purpose you mentioned, that Greg and Luke claim we can hotfix on the fly (if we can, great, let's do that, but if not...). it's SO obvious you cannot fork at the last minute. it's simply b/c a huge portion of the network won't get the memo to upgrade immediately. merchants will keep processing tx's that get forked off the network. chaos and confusion will occur. this is why Satoshi talked about a slow progressive "versioning" update with the final features only being enabled after 8-12 mo or so when it's clear thru monitoring that most everyone has updated. this is what we're doing right now with XT. i think it's really disingenuous for Greg to say so. yet another black mark in my checkbook for him. even if the network is lucky enough to get to the "last minute" before losing all it's customers from unconf tx's, tell me how you know you're there? apparently it will be decided by core devs like Greg. the obvious question is then, "where is the last minute?" if anything the last few weeks of contentious debate has demonstrated is that we can't decide on what the last minute will be even if we agree to wait until the last minute. I agree with you, they probably won't be able to do it. But as ZB said this proposal breaks the "1MB because that's the way its been" mentality even if it happens in 3 months instead of 1, so that is a good thing. Ironically, if it CAN be done in 1 month, that speaks pretty negatively about decentralization -- in the same sense that all the different central banks don't make the system decentralized. They are all marchin in step to the same drummer.
|
|
|
|
cypherdoc (OP)
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
June 04, 2015, 06:05:23 PM |
|
|
|
|
|
|
TPTB_need_war
|
|
June 04, 2015, 06:07:02 PM Last edit: June 04, 2015, 06:23:59 PM by TPTB_need_war |
|
if one's fundamental unit is the full node andis not the user, i think you're doing it wrong
I corrected that for you to stop violating the End-to-End Principle of networks. i doubt that any cryptocoin can ever be on auto pilot as the crypto evolves as computerization advances. what is secure today won't be secure tomorrow thus requiring continual updating.
The voters are the nodes, some developers want to keep centralized control over the majority of the nodes, some developers realize this is bad, - maintaining control its either explicit, subconscious, or subverted.
this is why Satoshi talked about a slow progressive "versioning" update with the final features only being enabled after 8-12 mo or so when it's clear thru monitoring that most everyone has updated
i think it's really disingenuous for Greg to say so
if it CAN be done in 1 month, that speaks pretty negatively about decentralization... They are all marchin in step to the same drummer.
An ideal crypto-coin would not violate Tim Berners-Lee's Principle of Least Power as Bitcoin egregiously does. It would do the minimum necessary and leave as much autonomy as possible to the nodes. Ideally the nodes could even disagree about the issues you all are squabbling about and the minimum requirement would still be met. It would be decentralized at any scale. It would scale to any level of transaction volume. It would not require any specific choice of crypto algorithm (nodes would be free to choose). Anyone guessed my paradigm shift yet?
|
|
|
|
cypherdoc (OP)
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
June 04, 2015, 06:07:30 PM |
|
looks where all the voters come from. does THIS look decentralized to you?:
|
|
|
|
Adrian-x
Legendary
Offline
Activity: 1372
Merit: 1000
|
|
June 04, 2015, 06:17:43 PM |
|
i doubt that any cryptocoin can ever be on auto pilot as the crypto evolves as computerization advances. what is secure today won't be secure tomorrow thus requiring continual updating.
The voters are the nodes, some developers want to keep centralized control over the majority of the nodes, some developers realize this is bad, - maintaining control its either explicit, subconscious, or subverted.
this is why Satoshi talked about a slow progressive "versioning" update with the final features only being enabled after 8-12 mo or so when it's clear thru monitoring that most everyone has updated
i think it's really disingenuous for Greg to say so
if it CAN be done in 1 month, that speaks pretty negatively about decentralization... They are all marchin in step to the same drummer.
An ideal crypto-coin would not violate Tim Berners-Lee's Principle of Least Power as Bitcoin egregiously does. It would do the minimum necessary and leave as much autonomy as possible to the nodes. Ideally the nodes could even disagree about the issues you all are squabbling about and the minimum requirement would still be met. It would be decentralized at any scale. It would scale to any level of transaction volume. It would not require any specific choice of crypto algorithm (nodes would be free to choose). Anyone guessed my paradigm shift yet? fiat 2.0, SDR's come on.
|
Thank me in Bits 12MwnzxtprG2mHm3rKdgi7NmJKCypsMMQw
|
|
|
TPTB_need_war
|
|
June 04, 2015, 06:21:25 PM Last edit: June 05, 2015, 06:05:40 AM by TPTB_need_war |
|
fiat 2.0, SDR's come on.
Hehe. Bitcoin is centralized, you are only obfuscating to yourself if you claim that it isn't. The centralization was put in the wrong place in Bitcoin's design. Move it then the decentralization can control the centralization.
|
|
|
|
|
Zangelbert Bingledack
Legendary
Offline
Activity: 1036
Merit: 1000
|
|
June 04, 2015, 06:41:01 PM |
|
if one's fundamental unit is the full node andis not the user, i think you're doing it wrong
I corrected that for you to stop violating the End-to-End Principle of networks. ... An ideal crypto-coin would not violate Tim Berners-Lee's Principle of Least Power as Bitcoin egregiously does. It would do the minimum necessary and leave as much autonomy as possible to the nodes. Ideally the nodes could even disagree about the issues you all are squabbling about and the minimum requirement would still be met. It would be decentralized at any scale. It would scale to any level of transaction volume. It would not require any specific choice of crypto algorithm (nodes would be free to choose). Anyone guessed my paradigm shift yet? How do you create a network topology that is decentralized at any scale?
|
|
|
|
NewLiberty
Legendary
Offline
Activity: 1204
Merit: 1002
Gresham's Lawyer
|
|
June 04, 2015, 06:43:26 PM |
|
Anyone guessed my paradigm shift yet?
The goal is clear enough, and laudable. It is the path to that goal that remains occluded.
|
|
|
|
Zangelbert Bingledack
Legendary
Offline
Activity: 1036
Merit: 1000
|
|
June 04, 2015, 06:43:39 PM |
|
Yes because almost everyone agrees it should be raised (Greg and Luke excepted, though they agree it should be raised eventually). The poll didn't ask about raising to 20MB.
|
|
|
|
NewLiberty
Legendary
Offline
Activity: 1204
Merit: 1002
Gresham's Lawyer
|
|
June 04, 2015, 06:47:33 PM |
|
Yes because almost everyone agrees it should be raised (Greg and Luke excepted, though they agree it should be raised eventually). The poll didn't ask about raising to 20MB.
The devil is in the details, with no details, its all angels. I suppose I would vote "NO" to most any change without knowing the details, but not "NO" to change itself. I think Gavin is getting closer and closer to something that will make sense. He is down to 8mb now, changed at a particular future block?
|
|
|
|
pinky
|
|
June 04, 2015, 06:56:26 PM |
|
looks where all the voters come from. does THIS look decentralized to you?: And what's the point of voting on highly technical question? Democracy doesn't work with this stuff, because majority have holes in their brains and then they lean on opinions of others and marketing campaigns.
|
|
|
|
impulse
|
|
June 04, 2015, 07:05:31 PM |
|
Interesting result. If that is truly reflective of the community as a whole then the anti-increase crown are quite disproportionately represented in this debate. Greg, Luke and Peter sure do know how to stir up some serious shit.
|
|
|
|
Zangelbert Bingledack
Legendary
Offline
Activity: 1036
Merit: 1000
|
|
June 04, 2015, 07:05:55 PM |
|
Ironically, if it CAN be done in 1 month, that speaks pretty negatively about decentralization -- in the same sense that all the different central banks don't make the system decentralized. They are all marchin in step to the same drummer.
I don't see how that's the case. If it's a good idea, and everyone's doing it, then the trade-off becomes attractive for just about everyone, so it can happen pretty fast without any centralization. Marching in lockstep doesn't mean centralized (of course; that's the whole idea of Bitcoin in a way). Isn't there some kind of alert key that goes to every full node? SPV nodes don't matter since they don't care about blocksize. If the problem is just alerting everyone that a fork is coming and they have to choose sides, then what is the worst that can happen? Can someone even operate a full node without seeing the alert?
|
|
|
|
NewLiberty
Legendary
Offline
Activity: 1204
Merit: 1002
Gresham's Lawyer
|
|
June 04, 2015, 07:09:08 PM |
|
Ironically, if it CAN be done in 1 month, that speaks pretty negatively about decentralization -- in the same sense that all the different central banks don't make the system decentralized. They are all marchin in step to the same drummer.
I don't see how that's the case. If it's a good idea, and everyone's doing it, then the trade-off becomes attractive for just about everyone, so it can happen pretty fast without any centralization. Marching in lockstep doesn't mean centralized (of course; that's the whole idea of Bitcoin in a way). Isn't there some kind of alert key that goes to every full node? SPV nodes don't matter. If the problem is just alerting everyone that a fork is coming and they have to choose sides, then what is the worst that can happen? Can someone even operate a full node without seeing the alert? This begs the question. It presupposes success. When a fork occurs, you are on one side or the other at the time of the fork. "If everyone's doing it" means that they have already done it, so there is no longer anything to be attractive to them. It is an event, not a migration. Unless you are imagining many many failed forks until one finally succeeds, which would be exceedingly chaotic.
|
|
|
|
|