Return-Path: Received: from smtp1.linuxfoundation.org (smtp1.linux-foundation.org [172.17.192.35]) by mail.linuxfoundation.org (Postfix) with ESMTPS id 52F9AACB for ; Fri, 26 Jun 2015 18:32:06 +0000 (UTC) X-Greylist: whitelisted by SQLgrey-1.7.6 Received: from mail-ig0-f169.google.com (mail-ig0-f169.google.com [209.85.213.169]) by smtp1.linuxfoundation.org (Postfix) with ESMTPS id 137FF1B8 for ; Fri, 26 Jun 2015 18:32:04 +0000 (UTC) Received: by igtg8 with SMTP id g8so7337937igt.0 for ; Fri, 26 Jun 2015 11:32:03 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc:content-type; bh=NhlHRDsu5fyVAbe48J9Nd3ltdP+qkuKbOXOgTcxogqQ=; b=ja/hn/3p3KC2+YfCRZa6Sa7VO/uzOQpqteYswlI83CGHiC6kE+GaJk5Ei97ZuPbeqY BszaUp95TkHyB3fHU//jvuDKGCd+2kRG/jd5P30zzudgJPvUqR5m3vN2Q5LAI7aFTjL6 rWfeBWGEqomevcxtsezempT9fbrNWT0LUbcMfVP6H3jXC+C4IbZeTSNENDhWjFdk3oZ1 17kquEGRM36RgfnOg6U+UFVTLNEwRkLzuLYFKBlttyOefQ2cW0KzCfA85ZlLsuXApAKX XyKLZCWyU/jpWNDC9e33ktXHDd4Sc6B/S9J+KzNxyH6EdUkw8xSyW1zSDKb05Q4bElGC 8Z2w== X-Gm-Message-State: ALoCoQkzAr0byuLzGb2+SM31zm2jvhDXHb3okhCuhHN8+f67msK3gwwYCOKuYbNSAzs8v2MFRU2C X-Received: by 10.107.37.69 with SMTP id l66mr3933458iol.76.1435343523483; Fri, 26 Jun 2015 11:32:03 -0700 (PDT) MIME-Version: 1.0 Received: by 10.107.149.20 with HTTP; Fri, 26 Jun 2015 11:31:43 -0700 (PDT) X-Originating-IP: [208.54.4.150] In-Reply-To: References: From: Mark Friedenbach Date: Fri, 26 Jun 2015 11:31:43 -0700 Message-ID: To: Jeff Garzik Content-Type: multipart/alternative; boundary=001a1140ea02d9307b05196ff4c9 X-Spam-Status: No, score=-2.6 required=5.0 tests=BAYES_00,HTML_MESSAGE, RCVD_IN_DNSWL_LOW autolearn=ham version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on smtp1.linux-foundation.org Cc: bitcoin-dev@lists.linuxfoundation.org Subject: Re: [bitcoin-dev] The need for larger blocks 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: Fri, 26 Jun 2015 18:32:06 -0000 --001a1140ea02d9307b05196ff4c9 Content-Type: text/plain; charset=UTF-8 Jeff, block size limits large enough to prevent fee pressure is absolutely, unequivocally unsustainable. We are already running against technological limits in the tradeoff between decentralization and utility. Increases of the block size limit in advance of fee pressure only delay the problem -- it does not and cannot solve it! We must be careful to use the block size limit now to get infrastructure to support a world with full blocks -- it's not that hard -- while still having a little room to grow fast if things unexpectedly break. On Fri, Jun 26, 2015 at 11:23 AM, Jeff Garzik wrote: > Failure to plan now for a hard fork increase 6(?) months in the future > produces that lumpy, unpredictable market behavior. > > The market has baked in the years-long behavior of low fees. From the > market PoV, inaction does lead to precisely that, a sudden change over the > span of a few months. > > At a higher level, people look at bitcoin and see people delaying, > waiting, dawdling until the barn is actually on fire before taking action > to put out the fire. > > They see a system that is not responsive to higher level externalities of > people & businesses making plans for the future. Based on current proposal > of change-through-inaction, businesses will simply shelve plans to use > bitcoin and not bother putting those new users on the network. > > If you wait until the need to increase block size is acute, it is already > too late. (1) Businesses have permanently shelved plans to use bitcoin and > (2) change at that point produces _larger_ disruption to the fee market. > > Hard forks require planning many months in advance. Gavin's timing is > sound, even though the Gavin/Hearn Bitcoin-XT antics were sub-optimal. > > > > > > > > On Fri, Jun 26, 2015 at 11:12 AM, Pieter Wuille > wrote: > >> I am not saying that economic change is what we want. Only that it is >> inevitable, independent of whether larger blocks happen or not. >> >> I am saying that acting because of fear of economic change is a bad >> reason. The reason for increase should be because of the higher utility. We >> need it at some point, but there should be no rush. >> >> I do understand that we want to avoid a *sudden* change in economic >> policy, but I'm generally not too worried. Either fees increase and they >> get paid, and we're good. But more likely is that some uses just move >> off-chain because the block chain does not offer what they need. That's >> sad, but it is inevitable at any size: some uses fit, some don't. >> >> -- >> Pieter >> On Jun 26, 2015 7:57 PM, "Jeff Garzik" wrote: >> >>> It is not "fear" of fee pressure. >>> >>> 1) Blocks are mostly not-full on average. >>> >>> 2) Absent long blocks and stress tests, there is little fee pressure >>> above the anti-spam relay fee metric, because of #1. >>> >>> 3) As such, inducing fee pressure is a delta, a change from years-long >>> bitcoin economic policy. Each time we approach the soft limit, Bitcoin >>> Core increases the soft limit to prevent "full" blocks. Mike Hearn et. al. >>> lobbies miners to upgrade. >>> >>> (note - this is not an endorsement of these actions - it is a neutral >>> observation) >>> >>> 4) Inaction leads to consistent fee pressure as the months tick on and >>> system volume grows; thus, inaction leads to economic policy change. >>> >>> 5) Economic policy change leads to market and software disruption. The >>> market and software - notably wallets - is not prepared for this. >>> >>> 6) If you want to change economic policy, that's fine. But be honest >>> and admit you are arguing for a change, a delta from current market >>> expectations and behavior. >>> >>> 7) It is critical to first deal with what _is_, not what you wish the >>> world to be. You want a fee market to develop. There is nothing wrong >>> with that desire. It remains a delta from where we are today, and that is >>> critically relevant in a $3b+ market. >>> >>> >>> >>> >>> >>> >>> >>> >>> On Fri, Jun 26, 2015 at 7:09 AM, Pieter Wuille >>> wrote: >>> >>>> Hello all, >>>> >>>> here I'm going to try to address a part of the block size debate which >>>> has been troubling me since the beginning: the reason why people seem to >>>> want it. >>>> >>>> People say that larger blocks are necessary. In the long term, I agree >>>> - in the sense that systems that do not evolve tend to be replaced by other >>>> systems. This evolution can come in terms of layers on top of Bitcoin's >>>> blockchain, in terms of the technology underlying various aspects of the >>>> blockchain itself, and also in the scale that this technology supports. >>>> >>>> I do, however, fundamentally disagree that a fear for a change in >>>> economics should be considered to necessitate larger blocks. If it is, and >>>> there is consensus that we should adapt to it, then there is effectively no >>>> limit going forward. This is similar to how Congress voting to increase the >>>> copyright term retroactively from time to time is really no different from >>>> having an infinite copyright term in the first place. This scares me. >>>> >>>> Here is how Gavin summarizes the future without increasing block sizes >>>> in PR 6341: >>>> >>>> > 1. Transaction confirmation times for transactions with a given fee >>>> will rise; very-low-fee transactions will fail to get confirmed at all. >>>> > 2. Average transaction fee paid will rise >>>> > 3. People or applications unwilling or unable to pay the rising fees >>>> will stop submitting transactions >>>> > 4. People and businesses will shelve plans to use Bitcoin, stunting >>>> growth and adoption >>>> >>>> Is it fair to summarize this as "Some use cases won't fit any more, >>>> people will decide to no longer use the blockchain for these purposes, and >>>> the fees will adapt."? >>>> >>>> I think that is already happening, and will happen at any scale. I >>>> believe demand for payments in general is nearly infinite, and only a small >>>> portion of it will eventually fit on a block chain (independent of whether >>>> its size is limited by consensus rules or economic or technological means). >>>> Furthermore, systems that compete with Bitcoin in this space already offer >>>> orders of magnitude more capacity than we can reasonably achieve with any >>>> blockchain technology at this point. >>>> >>>> I don't know what subset of use cases Bitcoin will cater to in the long >>>> term. They have already changed - you see way less betting transactions >>>> these days than a few years ago for example - and they will keep changing, >>>> independent of what effective block sizes we end up with. I don't think we >>>> should be afraid of this change or try to stop it. >>>> >>>> If you look at graphs of block sizes over time (for example, >>>> http://rusty.ozlabs.org/?p=498), it seems to me that there is very >>>> little "organic" growth, and a lot of sudden changes (which could >>>> correspond to changing defaults in miner software, introduction of popular >>>> sites/services, changes in the economy). I think these can be seen as the >>>> economy changing to full up the available space, and I believe these will >>>> keep happening at any size effectively available. >>>> >>>> None of this is a reason why the size can't increase. However, in my >>>> opinion, we should do it because we believe it increases utility and >>>> understand the risks; not because we're afraid of what might happen if we >>>> don't hurry up. And from that point of view, it seems silly to make a huge >>>> increase at once... >>>> >>>> -- >>>> Pieter >>>> >>>> >>>> _______________________________________________ >>>> 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 > > --001a1140ea02d9307b05196ff4c9 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Jeff, block size limits large enough to prevent fee p= ressure is absolutely, unequivocally unsustainable. We are already running = against technological limits in the tradeoff between decentralization and u= tility. Increases of the block size limit in advance of fee pressure only d= elay the problem -- it does not and cannot solve it!

We must b= e careful to use the block size limit now to get infrastructure to support = a world with full blocks -- it's not that hard -- while still having a = little room to grow fast if things unexpectedly break.

On Fri, Jun 26, 2015 at 11:23 AM, = Jeff Garzik <jgarzik@gmail.com> wrote:
Failure to plan now for a hard fork increase= 6(?) months in the future produces that lumpy, unpredictable market behavi= or.

The market has baked in the years-long behavior of l= ow fees.=C2=A0 From the market PoV, inaction does lead to precisely that, a= sudden change over the span of a few months.

At a= higher level, people look at bitcoin and see people delaying, waiting, daw= dling until the barn is actually on fire before taking action to put out th= e fire.

They see a system that is not responsive t= o higher level externalities of people & businesses making plans for th= e future.=C2=A0 Based on current proposal of change-through-inaction, busin= esses will simply shelve plans to use bitcoin and not bother putting those = new users on the network.

If you wait until the ne= ed to increase block size is acute, it is already too late. =C2=A0(1) Busin= esses have permanently shelved plans to use bitcoin and (2) change at that = point produces _larger_ disruption to the fee market.

<= div>Hard forks require planning many months in advance.=C2=A0 Gavin's t= iming is sound, even though the Gavin/Hearn Bitcoin-XT antics were sub-opti= mal.




=


On Fri, Jun 26, 2015 at 11:12 AM, Pieter Wuille <pieter.wuille@gmail.com> wrote:

I am not saying that economic change is what we wa= nt. Only that it is inevitable, independent of whether larger blocks happen= or not.

I am saying that acting because of fear of economic change i= s a bad reason. The reason for increase should be because of the higher uti= lity. We need it at some point, but there should be no rush.

I do understand that we want to avoid a *sudden* change in e= conomic policy, but I'm generally not too worried. Either fees increase= and they get paid, and we're good. But more likely is that some uses j= ust move off-chain because the block chain does not offer what they need. T= hat's sad, but it is inevitable at any size: some uses fit, some don= 9;t.

--
Pieter

On Jun 26, 2015 7:57 PM, "Jeff Garzik"= <jgarzik@gmail.c= om> wrote:
It is not "fear" of fee pressure.

1) Blocks are mostly not-full on average.

2= ) Absent long blocks and stress tests, there is little fee pressure above t= he anti-spam relay fee metric, because of #1.

3) A= s such, inducing fee pressure is a delta, a change from years-long bitcoin = economic policy.=C2=A0 Each time we approach the soft limit, Bitcoin Core i= ncreases the soft limit to prevent "full" blocks.=C2=A0 Mike Hear= n et. al. lobbies miners to upgrade.

(note - this = is not an endorsement of these actions - it is a neutral observation)
=

4) Inaction leads to consistent fee pressure as the mon= ths tick on and system volume grows; thus, inaction leads to economic polic= y change.

5) Economic policy change leads to marke= t and software disruption.=C2=A0 The market and software - notably wallets = - is not prepared for this.

6) If you want to chan= ge economic policy, that's fine.=C2=A0 But be honest and admit you are = arguing for a change, a delta from current market expectations and behavior= .

7) It is critical to first deal with what _is_, = not what you wish the world to be.=C2=A0 You want a fee market to develop.= =C2=A0 There is nothing wrong with that desire.=C2=A0 It remains a delta fr= om where we are today, and that is critically relevant in a $3b+ market.








On Fri, Jun 26, 2015 at 7:09 AM, Pieter Wuille <pieter.wuille@gmail.com> wrote:
Hello = all,

here I'm going to try to address a part of the block size debate which has been=20 troubling me since the beginning: the reason why people seem to want it.
People say that larger blocks are necessary. In the long term, I agree - in=20 the sense that systems that do not evolve tend to be replaced by other=20 systems. This evolution can come in terms of layers on top of Bitcoin's= =20 blockchain, in terms of the technology underlying various aspects of the blockchain itself, and also in the scale that this technology supports.
I do, however, fundamentally disagree that a fear for a change in=20 economics should be considered to necessitate larger blocks. If it is,=20 and there is consensus that we should adapt to it, then there is=20 effectively no limit going forward. This is similar to how Congress=20 voting to increase the copyright term retroactively from time to time is really no different from having an infinite copyright term in the first place. This scares me.

Here is how Gavin summarizes the futur= e without increasing block sizes in PR 6341:

> 1. Transaction con= firmation times for transactions with a given fee=20 will rise; very-low-fee transactions will fail to get confirmed at all.
= > 2. Average transaction fee paid will rise
> 3. People or applica= tions unwilling or unable to pay the rising fees will stop submitting trans= actions
> 4. People and businesses will shelve plans to use Bitcoin, = stunting growth and adoption

Is it fair to summarize this as "Some use cases won't fit any more, = people will decide to no longer use the blockchain for these purposes, and the fees will adapt."?

I think that is already happening, an= d=20 will happen at any scale. I believe demand for payments in general is=20 nearly infinite, and only a small portion of it will eventually fit on a block chain (independent of whether its size is limited by consensus=20 rules or economic or technological means). Furthermore, systems that=20 compete with Bitcoin in this space already offer orders of magnitude=20 more capacity than we can reasonably achieve with any blockchain=20 technology at this point.

I don't know what subset of use cases= =20 Bitcoin will cater to in the long term. They have already changed - you=20 see way less betting transactions these days than a few years ago for=20 example - and they will keep changing, independent of what effective=20 block sizes we end up with. I don't think we should be afraid of this= =20 change or try to stop it.

If you look at graphs of block sizes= over time (for example, http://rusty.ozlabs.org/?p=3D498), it seems to me that there is very little "organic" growth, and a= lot of sudden changes (which could correspond to changing defaults in miner=20 software, introduction of popular sites/services, changes in the=20 economy). I think these can be seen as the economy changing to full up=20 the available space, and I believe these will keep happening at any size effectively available.

None of this is a reason why the=20 size can't increase. However, in my opinion, we should do it because we= =20 believe it increases utility and understand the risks; not because we'r= e afraid of what might happen if we don't hurry up. And from that point= =20 of view, it seems silly to make a huge increase at once...

--
Pieter


_______________________________________________
bitcoin-dev mailing list
= bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mail= man/listinfo/bitcoin-dev




_______________________________________________
bitcoin-dev mailing list
= bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mail= man/listinfo/bitcoin-dev


--001a1140ea02d9307b05196ff4c9--