Date: Mon, 25 Apr 2022 10:01:28 +0000
To: Zac Greenwood <zachgrw@gmail.com>
From: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Reply-To: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Message-ID: <ID4VQ7kbkdRwf38kiXtxuhanlkFy7ggmL26NlK_t4h9KdAzrIMFXGdgwsGxIixQXhkdyeaSupmVEVDp14aQPwShHm1AryWFiYAkZO5Thb4k=@protonmail.com>
In-Reply-To: <CAJ4-pEB1zWxWapt3qit=X1p3WuSSn2_K=rygDp8JzpS7GimgMg@mail.gmail.com>
References: <RyYBRY3MJP_0b2YkCEUFBdP8u1A_cGSEEkDbzKK9k-rkINZrBaOL70L96iHR11bJhmkhAzuN6uZ1X8PQgz2wa8Us3-2OpNa4RbhSSprw_WE=@protonmail.com>
 <CALeFGL1=4PrA_ziTsoS9sUjGjfLr54AiMfM99uDV-Bau5Ab_eQ@mail.gmail.com>
 <CAJ4-pEADrHf_YR5ZBfJW+eefKrp1iEj4wAi72UrwRSi9gaVP+w@mail.gmail.com>
 <B037C028-EB25-4197-A0F1-48838D1AD30F@petertodd.org>
 <tBEO_7gKqAlgiRpjC1OCLMBcLddK0ffNlbdU65VhXnXqZQiDCTDNOw8sfTXqNXZxVA8wA0uodTLvptt7Iy5O3D_6hkiVVfLegOrg9C1Snkg=@protonmail.com>
 <CAJ4-pEB1zWxWapt3qit=X1p3WuSSn2_K=rygDp8JzpS7GimgMg@mail.gmail.com>
Feedback-ID: 2872618:user:proton
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>
Subject: Re: [bitcoin-dev] User Resisted Soft Fork for CTV
Precedence: list

Good morning Zac,

> On Mon, 25 Apr 2022 at 07:36, ZmnSCPxj <zmnscpxj@protonmail.com> wrote
>
> > CTV *can* benefit layer 2 users, which is why I switched from vaguely a=
pathetic to CTV, to vaguely supportive of it.
>
>
> Other proposals exist that also benefit L2 solutions. What makes you supp=
ort CTV specifically?

It is simple to implement, and a pure `OP_CTV` SCRIPT on a P2WSH / P2SH is =
only 32 bytes + change on the output and 32 bytes + change on the input/wit=
ness, compared to signature-based schemes which require at least 32 bytes +=
 change on the output and 64 bytes + change on the witness ***IF*** they us=
e the Taproot format (and since we currently gate the Taproot format behind=
 actual Taproot usages, any special SCRIPT that uses Taproot-format signatu=
res would need at least the 33-byte internal pubkey revelation; if we settl=
e with the old signature format, then that is 73 bytes for the signature).
To my knowledge as well, hashes (like `OP_CTV` uses) are CPU-cheaper (and m=
emory-cheaper?) than even highly-optimized `libsecp256k1` signature validat=
ion, and (to my knowledge) you cannot use batch validation for SCRIPT-based=
 signature checks.
It definitely does not enable recursive covenants, which I think deserve mo=
re general research and thinking before we enable recursive covenants.

Conceptually, I see `OP_CTV` as the "AND" to the "OR" of MAST.
In both cases, you have a hash-based tree, but in `OP_CTV` you want *all* t=
hese pre-agreed cases, while in MAST you want *one* of these pre-agreed cas=
es.

Which is not to say that other proposals do not benefit L2 solutions *more*=
 (`SIGHASH_ANYPREVOUT` when please?), but other proposals are signature-bas=
ed and would be larger in this niche.

Regards,
ZmnSCPxj