[bitcoin-dev] Thoughts on soft-fork activation

ZmnSCPxj ZmnSCPxj at protonmail.com
Fri Jul 17 02:58:46 UTC 2020


Good morning list, BlueMatt and aj,

There is an idea circulating on IRC and elsewhere, which seems to be at least 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=true` `lockin=+42 months` (or 36 months, 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 good 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=true` that will have a timeout that ends +42 months after the first countdown, even with Modern Softfork Activation.
* The abort is roughly equivalent to the Modern Softfork Activation case where during the 6 month discussion period we decide not to deploy Taproot after all.
* The deployment of a BIP91 reduced-threshold 80% approximates what aj proposes, to reduce the threshold for activation later.

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


Regards,
ZmnSCPxj


More information about the bitcoin-dev mailing list