Date: Fri, 17 Jul 2020 02:58:46 +0000
To: Matt Corallo <lf-lists@mattcorallo.com>,
 Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
From: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Reply-To: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Message-ID: <T1HhFz-L_-xgdFMT7fCoP5LV_eigU8Kn8pwrWzevW13vX-wX-KvslKVWHo7IJQRYh2I4w4RsnLpl5gF6F6FCDL_6Gy80F8NOhO38vKqIu0Q=@protonmail.com>
In-Reply-To: <9c401353-d63a-3d19-72a8-ffcaf169ac6d@mattcorallo.com>
References: <20200714093730.myvls2jfpwyi3ap3@erisian.com.au>
 <9c401353-d63a-3d19-72a8-ffcaf169ac6d@mattcorallo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [bitcoin-dev] Thoughts on soft-fork activation
Precedence: list

Good morning list, BlueMatt and aj,

There is an idea circulating on IRC and elsewhere, which seems to be at lea=
st mildly supported by gmax and roconnor, which I will try to explain here.

(These are my words, so if there is some mistake, I apologize)

Basically:

* Deploy a BIP8 `lockinontimeout=3Dtrue` `lockin=3D+42 months` (or 36 month=
s, or 24 months) at next release.
  * Pedanty note: BIP8 uses blockheights, not actual times.
* Then 1 year after `starttime`, ***if*** it is not activated yet:
  * Discuss.
  * If we think it is only because of miner apathy and user support seems g=
ood regardless, deploy a BIP91 reduced-threshold 80% that enforces the BIP8=
 bit.
    * We hope that this will stave off independent attempts at a UASF with =
a faster timeout.
  * If we think there are real reasons not to continue with Taproot as-is, =
deploy an abort: a softfork that disallows transaction outputs with `OP_1 <=
32-bytes>` `scriptPubKey` (other lengths and other versions are allowed).

This approximates what aj is proposing:

* Ultimately, we expect to deploy a BIP8 `lockinontimeout=3Dtrue` that will=
 have a timeout that ends +42 months after the first countdown, even with M=
odern Softfork Activation.
* The abort is roughly equivalent to the Modern Softfork Activation case wh=
ere during the 6 month discussion period we decide not to deploy Taproot af=
ter all.
* The deployment of a BIP91 reduced-threshold 80% approximates what aj prop=
oses, to reduce the threshold for activation later.

As I understand it, an advantage of this proposal is that we can deploy ver=
y quickly a relatively simple BIP8 `locktimeontimeout=3Dtrue`, then continu=
e debate on various details (is 80% too low? too high? are users actually d=
eploying? are mining pools updating? etc) in parallel.
This lets the code into the hands of users where they can start deploying i=
t and we can start getting better gauges on how well Taproot is supported.


Regards,
ZmnSCPxj