MIME-Version: 1.0
In-Reply-To: <4e57154a051f182bf9c4f939a61acad3@xbt.hk>
References: <d17549688c0c747b2077c1f6f96b6445@xbt.hk>
	<CABm2gDrMome0xZGvPTYr9DaFJt=Si0Lmv=VTa4ydd4Bj6ARTcw@mail.gmail.com>
	<4e57154a051f182bf9c4f939a61acad3@xbt.hk>
Date: Wed, 19 Aug 2015 12:53:30 +0200
Message-ID: <CABm2gDpce_5CCeAGFqqDZvkC6aV9p-cU15OrG4PjMODFBKY45Q@mail.gmail.com>
From: =?UTF-8?B?Sm9yZ2UgVGltw7Nu?= <jtimon@jtimon.cc>
To: jl2012@xbt.hk
Content-Type: text/plain; charset=UTF-8
Cc: Bitcoin Dev <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] Bitcoin is an experiment. Why don't we have an
 experimental hardfork?
Precedence: list

On Wed, Aug 19, 2015 at 12:34 PM,  <jl2012@xbt.hk> wrote:
> You misunderstand my intention. The experiment is not about a random
> hardfork. It's about a block size increase hardfork.

One of your goals is "show the world that reaching consensus for a
Bitcoin hardfork is possible", right?
BIP99 can achieve that goal without touching the block size (thus
probably less controversial).

> The data will help us
> to design further hardfork on block size.

The data can be collected using testchains. See #6382

> To make it less controversial, the size must not be too big.

That makes the data less interesting, doesn't it?

> To allow a meaningful experiment, the size must not be too small.
> Technically we could make it 1.01MB but that defeats all objectives I listed
> and there is no point to do it.

It wouldn't defeat the "show the world that reaching consensus for a
Bitcoin hardfork is possible" objective though. But again, we don't
need to touch the block size to achieve that goal.

> That's why I suggest 1.5MB.

But that wouldn't be uncontroversial (unless it was accompanied with
data that somehow quantifies the risks, in which case maybe another
bigger size is acceptable).

>>> 1. Today, we all agree that some kind of block size hardfork will happen
>>> on
>>> t1=*1 June 2016*
>>
>>
>> I disagree with this. I think it should be schedule at least a year
>> after it is deployed in the newest versions.
>> Maybe there's something special about June 2016 that I'm missing.
>
>
> I hope the fork could be done before the halving, which (hopefully) we may
> have a new bitcoin rush
>
> There was only 2 months for the BIP50 hardfork. You may argue that's a "bug
> fix" but practically there is no difference: people not fixing the bug in 2
> months was forked off. Four months of grace period (Feb to June 2016) is
> already a double of that.

Yes, emergency hardforks are a different case as explained in BIP99's draft.
In any case, " we all agree that some kind of block size hardfork will
happen on June 2016" it's clearly false since I can find many
counter-examples besides myself.

> Also, if we could have zero grace period for softfork, why must we have a
> ultra-long period for hardfork? (Unless you also agree to have an 1-year
> grace period for softfork. I don't buy the "softfork is safer than hardfork"
> theory. The recent BIP66 fork has clearly shown why it is wrong:
> non-upgrading full nodes are not full nodes)

I don't think giving 1 year for "everybody in the world" to upgrade is
an "ultra-long" period.
I'm proposing 5 years for the hardfork proposed in bip99.
I do think softforks are safer than hardforks, that doesn't mean
softforks don't have serious risks as well.

> The problem is many people won't update until they must do so. So 4 months
> or 1 year make little difference

I disagree with this conclusion as well (it doesn't follow from the
first sentence, non sequitur).