Date: Fri, 18 Feb 2022 14:48:38 +0000
To: Erik Aronesty <erik@q32.com>
From: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Reply-To: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Message-ID: <mSBTc8Bl5YIXe7LSX_fCNUYhd0wjepa_XhF6uhtwzF7s5h9-AEGWbkfrPA58nn431SjAqTkWEzd7YJ5mC0M7aZf-NmS5eDTN8LKEGQOFGcY=@protonmail.com>
In-Reply-To: <CAJowKg+cK3ZjPCjcDK8v5qFA=uCHD7gcR8ymroXBFicU5jzY8Q@mail.gmail.com>
References: <6nZ-SkxvJLrOCOIdUtLOsdnl94DoX_NHY0uwZ7sw78t24FQ33QJlJU95W7Sk1ja5EFic5a3yql14MLmSAYFZvLGBS4lDUJfr8ut9hdB7GD4=@protonmail.com>
 <CAJowKg+cK3ZjPCjcDK8v5qFA=uCHD7gcR8ymroXBFicU5jzY8Q@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Cc: Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>,
 Anthony Towns <aj@erisian.com.au>
Subject: Re: [bitcoin-dev] `OP_EVICT`: An Alternative to
	`OP_TAPLEAFUPDATEVERIFY`
Precedence: list

Good morning Erik,

> hey, i read that whole thing, but i'm confused as to why it's necessary
>
> seems like N of N participants can pre-sign an on-chain transfer of funds=
 for each participant to a new address that consists of (N-1) or (N-1) part=
icipants, of which each portion of the signature is encrypted for the same =
(N-1) participants
>
> then any (N-1) subset of participants can collude publish that transactio=
n at any time to remove any other member from=C2=A0the pool
>
> all of the set up=C2=A0 (dkg for N-1), and transfer (encryption of partia=
l sigs) is done offchain, and online with the participants=C2=A0that are on=
line


As I understand your counterproposal, it would require publishing one trans=
action per evicted participant.
In addition, each participant has to store `N!` possible orderings in which=
 participants can be evicted, as you cannot predict the future and cannot p=
redict which partiicpants will go offline first.

Finally, please see also the other thread on lightning-dev: https://lists.l=
inuxfoundation.org/pipermail/lightning-dev/2022-February/003479.html
In this thread, I point out that if we ever use channel factories, it would=
 be best if we treat each channel as a 2-of-2 that participates in an overa=
ll N-of-N (i.e. the N in the outer channel factory is composed of 2-of-2).
For example, instead of the channel factory being signed by participants `A=
`, `B`, `C`, `D`, instead the channel factory is signed by `AB`, `AC`, `AD`=
, `BC`, `BD`, `CD`, so that if e.g. participant B needs to be evicted, we c=
an evict the signers `AB`, `BC`, and `BD`.
This means that for the channel factory case, already the number of "partic=
ipants" is quadratic on the number of *actual* participants, which greatly =
increases the number of transactions that need to be evicted in one-evictio=
n-at-a-time schemes (which is how I understand your proposal) as well as in=
creasing the `N!` number of signatures that need to be exchanged during set=
up.


But yes, certainly that can work, just as pre-signed transactions can be us=
ed instead of `OP_CTV` or pretty much any non-`OP_CHECKMULTISIG` opcode, xr=
ef Smart Contracts Unchained.

Regards,
ZmnSCPxj