Return-Path: Received: from smtp1.linuxfoundation.org (smtp1.linux-foundation.org [172.17.192.35]) by mail.linuxfoundation.org (Postfix) with ESMTPS id 087F983D for ; Thu, 25 Jun 2015 02:34:53 +0000 (UTC) X-Greylist: from auto-whitelisted by SQLgrey-1.7.6 Received: from mail.help.org (mail.help.org [70.90.2.18]) by smtp1.linuxfoundation.org (Postfix) with ESMTPS id 861D9180 for ; Thu, 25 Jun 2015 02:34:51 +0000 (UTC) Received: from [10.1.10.25] (B [10.1.10.25]) by mail.help.org with ESMTPA ; Wed, 24 Jun 2015 22:34:49 -0400 To: bitcoin-dev@lists.linuxfoundation.org References: <558B4632.8080504@bitcoins.info> From: Milly Bitcoin Message-ID: <558B68C3.9050608@bitcoins.info> Date: Wed, 24 Jun 2015 22:34:43 -0400 User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:38.0) Gecko/20100101 Thunderbird/38.0.1 MIME-Version: 1.0 In-Reply-To: Content-Type: multipart/alternative; boundary="------------060105070409060504020206" X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00,HTML_MESSAGE autolearn=ham version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on smtp1.linux-foundation.org Subject: Re: [bitcoin-dev] BIP Process and Votes X-BeenThere: bitcoin-dev@lists.linuxfoundation.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: Bitcoin Development Discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 25 Jun 2015 02:34:53 -0000 This is a multi-part message in MIME format. --------------060105070409060504020206 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit I'm sorry but that is the kind of defensive, cultish response everyone gets when they ask that question. If you had a well constructed documented process then you would be able to point to it ... but you can't. While there are a few bits and pieces scattered about in different places there is no coherent plan or process. It is easy to make statements like "consensus must be unanimous" but the issue is that you never have true 100% consensus yet you have to move forward in some fashion and everyone has to run software with the same consensus rules. The issue is how you move forward is the question that nobody wants to answer because (a) it is a hard question to answer and (b) developers see it as a threat to their authority/position. If people just keep shutting down the discussion with a bunch of cultish stock answers then you are never going to move forward with developing some kind of process. From what I can see much of the discussion is personality-driven and not based on Computer Science or and defined process. The issue is that a personality has changed so the process is perceived to be different and some people want to hard fork. Previously, the cultish answer is that Bitcoin development is decentralized because people can fork the code. Now that some developers want to fork the code suddenly it is a big problem. Is forking the code part of the consensus process or is it the work of the devil? The fact that there is so much diverse opinion on this shows a defined process has never been fully vetted or understood. I have worked on these processes for many years for projects orders of magnitudes larger than Bitcoin. I can absolutely assure you the current mishmash does not scale and huge amounts of time are wasted. That should be readily apparent from the recent discussions and the recent concern it has caused from people outside the developer's inner circle. Lack of defined process = high risk and wasted effort. Russ On 6/24/2015 9:50 PM, Mark Friedenbach wrote: > I'm sorry but this is absolutely not the case, Milly. The reason that > people get defensive is that we have a carefully constructed process > that does work (thank you very much!) and is well documented. We talk > about it quite often in fact as it is a defining characteristic of how > bitcoin is developed which differs in some ways from how other open > source software is developed -- although it remains the same in most > other ways. > > Changes to the non-consensus sections of Bitcoin Core tend to get > merged when there are a few reviews, tests, and ACKs from recognized > developers, there are no outstanding objections, and the maintainer > doing the merge makes a subjective judgement that the code is ready. > > Consensus-changes, on the other hand, get merged into Bitcoin Core > only after the above criteria are met AND an extremely long discussion > period that has given all the relevant stakeholders a chance to > comment, and no significant objections remain. Consensus-code changes > are unanimous. They must be. > > The sort of process that exists in standards bodies for example, with > working groups and formal voting procedures, has no place where > changes define the nature and validity of other people's money. Who > has the right to reach into your pocket and define how you can or > cannot spend your coins? The premise of bitcoin is that no one has > that right, yet that is very much what we do when consensus code > changes are made. That is why when we make a change to the rules > governing the nature of bitcoin, we must make sure that everyone is > made aware of the change and consents to it. > > Everyone. Does this work? Does this scale? So far, it does. > Uncontroversial changes, such as BIP 66, are deployed without issue. > Every indication is that BIP 66 will complete deployment in the very > near future, and we intend to repeat this process for more interesting > changes such as BIP65: CHECKLOCKTIMEVERIFY. > > This isn't about no one stepping forward to be the "decider." This is > about no one having the right to decide these things on the behalf of > others. If a contentious change is proposed and not accepted by the > process of consensus, that is because the process is doing its job at > rejecting controversial changes. It has nothing to do with > personality, and everything to do with the nature of bitcoin itself. > > > On Wed, Jun 24, 2015 at 5:07 PM, Milly Bitcoin > wrote: > > I have seen this question asked many times. Most developers > become defensive and they usually give a very vague 1-sentence > answer when this question is asked. It seems to be it is based on > personalities rather than any kind of definable process. To have > that discussion the personalities must be separated out and > answers like "such-and-such wouldn't do that" don't really do much > to advance the discussion. Also, the incentive for new developers > to come in is that they will be paid by companies who want to > influence the code and this should be considered (some developers > take this statement as an insult when it is just a statement of > the incentive process). > > The other problem you are having is the lead developer does not > want to be a "decider" when, in fact, he is a very significant > decider. While the users have the ultimate choice in a practical > sense the chief developer is the "decider." Now people don't want > to get him upset so nobody wants to push the issue or fully define > the process. Now you are left with a broken, unwritten/unspoken > process. While this type of thing may work with a small group of > developers businesses/investors looking in from the outside will > see this as a risk. > > Until you get passed all the personality-based arguments you are > going to have a tough time defining a real process. > > Russ > > > > > > > On 6/24/2015 7:41 PM, Raystonn wrote: > > I would like to start a civil discussion on an undefined, or > at least unwritten, portion of the BIP process. Who should > get to vote on approval to commit a BIP implementation into > Bitcoin Core? Is a simple majority of these voters sufficient > for approval? If not, then what is? > > Raystonn > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists.linuxfoundation.org > > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > > > > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists.linuxfoundation.org > > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > > --------------060105070409060504020206 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: 8bit
I'm sorry but that is the kind of defensive, cultish response everyone gets when they ask that question.  If you had a well constructed documented process then you would be able to point to it ... but you can't.  While there are a few bits and pieces scattered  about in different places there is no coherent plan or process.

It is easy to make statements like "consensus must be unanimous" but the issue is that you never have true 100% consensus yet you have to move forward in some fashion and everyone has to run software with the same consensus rules.  The issue is how you move forward is the question that nobody wants to answer because (a) it is a hard question to answer and (b) developers see it as a threat to their authority/position.  If people just keep shutting down the discussion with a bunch of cultish stock answers then you are never going to move forward with developing some kind of process. 

From what I can see much of the discussion is personality-driven and not based on Computer Science or and defined process.  The issue is that a personality has changed so the process is perceived to be different and some people want to hard fork.  Previously, the cultish answer is that Bitcoin development is decentralized because people can fork the code.  Now that some developers want to fork the code suddenly it is a big problem.   Is forking the code part of the consensus process or is it the work of the devil?   The fact that there is so much diverse opinion on this shows a defined process has never been fully vetted or understood.

I have worked on these processes for many years for projects orders of magnitudes larger than Bitcoin.  I can absolutely assure you the current mishmash does not scale and huge amounts of time are wasted.  That should be readily apparent from the recent discussions and the recent concern it has caused from people outside the developer's inner circle. 

Lack of defined process = high risk and wasted effort.

Russ




On 6/24/2015 9:50 PM, Mark Friedenbach wrote:
I'm sorry but this is absolutely not the case, Milly. The reason that people get defensive is that we have a carefully constructed process that does work (thank you very much!) and is well documented. We talk about it quite often in fact as it is a defining characteristic of how bitcoin is developed which differs in some ways from how other open source software is developed -- although it remains the same in most other ways.

Changes to the non-consensus sections of Bitcoin Core tend to get merged when there are a few reviews, tests, and ACKs from recognized developers, there are no outstanding objections, and the maintainer doing the merge makes a subjective judgement that the code is ready.

Consensus-changes, on the other hand, get merged into Bitcoin Core only after the above criteria are met AND an extremely long discussion period that has given all the relevant stakeholders a chance to comment, and no significant objections remain. Consensus-code changes are unanimous. They must be.

The sort of process that exists in standards bodies for example, with working groups and formal voting procedures, has no place where changes define the nature and validity of other people's money. Who has the right to reach into your pocket and define how you can or cannot spend your coins? The premise of bitcoin is that no one has that right, yet that is very much what we do when consensus code changes are made. That is why when we make a change to the rules governing the nature of bitcoin, we must make sure that everyone is made aware of the change and consents to it.

Everyone. Does this work? Does this scale? So far, it does. Uncontroversial changes, such as BIP 66, are deployed without issue. Every indication is that BIP 66 will complete deployment in the very near future, and we intend to repeat this process for more interesting changes such as BIP65: CHECKLOCKTIMEVERIFY.

This isn't about no one stepping forward to be the "decider." This is about no one having the right to decide these things on the behalf of others. If a contentious change is proposed and not accepted by the process of consensus, that is because the process is doing its job at rejecting controversial changes. It has nothing to do with personality, and everything to do with the nature of bitcoin itself.


On Wed, Jun 24, 2015 at 5:07 PM, Milly Bitcoin <milly@bitcoins.info> wrote:
I have seen this question asked many times.  Most developers become defensive and they usually give a very vague 1-sentence answer when this question is asked.  It seems to be it is based on personalities rather than any kind of definable process.  To have that discussion the personalities must be separated out and answers like "such-and-such wouldn't do that" don't really do much to advance the discussion.  Also, the incentive for new developers to come in is that they will be paid by companies who want to influence the code and this should be considered (some developers take this statement as an insult when it is just a statement of the incentive process).

The other problem you are having is the lead developer does not want to be a "decider" when, in fact, he is a very significant decider.  While the users have the ultimate choice in a practical sense the chief developer is the "decider."  Now people don't want to get him upset so nobody wants to push the issue or fully define the process.  Now you are left with a broken, unwritten/unspoken process.  While this type of thing may work with a small group of developers businesses/investors looking in from the outside will see this as a risk.

Until you get passed all the personality-based arguments you are going to have a tough time defining a real process.

Russ






On 6/24/2015 7:41 PM, Raystonn wrote:
I would like to start a civil discussion on an undefined, or at least unwritten, portion of the BIP process.  Who should get to vote on approval to commit a BIP implementation into Bitcoin Core?  Is a simple majority of these voters sufficient for approval?  If not, then what is?

Raystonn
_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev



_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev


--------------060105070409060504020206--