Content-Type: text/plain;
	charset=utf-8
Mime-Version: 1.0 (1.0)
From: Eric Voskuil <eric@voskuil.org>
In-Reply-To: <B853EDF2-8A8A-44B0-A66E-F86175E61EDA@voskuil.org>
Date: Fri, 5 Jul 2019 16:20:17 -0700
Content-Transfer-Encoding: quoted-printable
Message-Id: <F9B77954-7F2A-4A8D-8ABF-502A9F48F503@voskuil.org>
References: <0DBC0DEA-C999-4AEE-B2E1-D5337ECD9405@gmail.com>
	<A4A6099F-F115-4CBF-B7D5-F16581476126@gmail.com>
	<063D7C06-F5D8-425B-80CE-CAE03A1AAD0C@voskuil.org>
	<0AA10217-E1CC-46D1-9B43-038CEEF942CD@gmail.com>
	<E72C4A8E-F850-400B-B19B-7D06B8A169EC@voskuil.org>
	<A64C3DCB-10CE-45EA-9A1B-7E3D13DF90EA@gmail.com>
	<6B9A04E2-8EEE-40A0-8B39-64AA0F478CAB@voskuil.org>
	<SEQmsx6ck79biVthBbBk1b9r9-R45sBwqWrv3FewQIBl4J18sOlwAPRt8sbTIbrBB8DX538GfwQkU40lyODmEkGSwah_VmbXT8iOr2Jcjlw=@protonmail.com>
	<F17F2E86-BFA4-456E-85F9-0D6583692AEC@voskuil.org>
	<kSCa9KUmpJox2_aglqhel-WdGlXf14mfKNZ95T4xqsrkQJ2Zh5zFA-Llq-j9cXX87iEPP5_aCkO9oR5kfQGKMBK9ds3Jct1V1FAawwa4CyE=@protonmail.com>
	<B853EDF2-8A8A-44B0-A66E-F86175E61EDA@voskuil.org>
To: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Cc: Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] Generalized covenants with taproot enable
	riskless or risky lending,
	prevent credit inflation through fractional reserve
Precedence: list


> On Jul 5, 2019, at 12:27, Eric Voskuil <eric@voskuil.org> wrote:
>=20
>=20
>> On Jul 4, 2019, at 21:05, ZmnSCPxj <ZmnSCPxj@protonmail.com> wrote:
>>=20
>> Good morning Eric,
>>=20
>>> As with Bitcoin mining, it is the consumed cost that matters in this sce=
nario, (i.e., not the hash rate, or in this case the encumbered coin face va=
lue). Why would the advertiser not simply be required to burn .1 coin for th=
e same privilege, just as miners burn energy? Why would it not make more sen=
se to spend that coin in support of the secondary network (e.g. paying for c=
onfirmation security), just as with the burning of energy in Bitcoin mining?=

>=20
> Good morning ZmnSCPxj,
>=20
>> Using the unspentness-time of a UTXO allows for someone advertising a ser=
vice or producer to "close up shop" by simply spending the advertising UTXO.=

>> For instance, if the advertisement is for sale of a limited stock of good=
s, once the stock has been sold, the merchant (assuming the merchant used ow=
n funds) can simply recover the locked funds, with the potential to reinvest=
 them elsewhere.
>> This allows some time-based hedging for the merchant (they may be willing=
 to wait indefinitely for the stock to be sold, but once the stock is sold, t=
hey can immediately reap the rewards of not having their funds locked anymor=
e).
>=20
> This is a materially different concept than proposed by Tamas.
>=20
> =E2=80=9C...he gives up his control of the coins until maturity, he can no=
t use them elsewhere until then.=E2=80=9D
>=20
>> Similarly, an entity renting out a UTXO for an advertisement might allow f=
or early reclamation of the UTXO in exchange for partial refund of fee; as t=
he value in the UTXO is now freed to be spent elsewhere, the lessor can leas=
e it to another advertiser.
>=20
> You appear to be proposing a design whereby either the owner or the renter=
 (not entirely clear to me which) can spend the =E2=80=9Clocked up=E2=80=9D c=
oin at any time (no maturity constraint), by dropping the covenant.
>=20
> If the renter can do this he can simply steal the coin from the owner.
>=20
> If the owner can do this there is no value to the renter (or as a proof of=
 cost), as the owner retains full control of the coin.
>=20
> If you mean that the age of the encumbrance is the proof of cost, this req=
uires no covenant. I don=E2=80=99t believe this is what you intended, just c=
overing all bases.
>=20
>> Burnt funds cannot be "un-burnt" to easily signal the end of a term for a=
n advertisement.
>=20
> And as I have shown above, nor can a =E2=80=9Clocked-up=E2=80=9D coin be u=
nlocked to do the same.
>=20
>> Similarly for miner fees.
>=20
> Well that=E2=80=99s the point, money spent is no longer under one=E2=80=99=
s control. The provable cost of this surrender was your stated objective. Re=
nting at a fractional cost of coin face value is a non-recoverable spend by t=
he renter to the owner. Burning or spending the same amount in a way that is=
 provably not to one=E2=80=99s self achieves the exact same result.
>=20
>> The best that can be done would be to have the nodes of the classified ad=
s network automatically decay the spent value of older advertisements to let=
 them be dropped from their advertisements pool.
>=20
> The advertiser can presumably trade control of as space on the ad network.=
 It=E2=80=99s not clear to me why this is not simply an independent chain of=
 limited ad space ownership. It might as well be namecoin.
>=20
>> Less importantly, burning currently has bad resource usage for practical a=
pplications.
>> Practical burning requires spending to a provably-unspendable P2PKH or P2=
SH or similar output.
>> This adds UTXO entries to the UTXO database that will never be removed.

I forgot to add that it is certainly possible to burn using a nonstandard sc=
ript, such as the non-zero OP_RETURN you suggested, without a consensus chan=
ge. This can be, as you say, made more practical with a policy change. But s=
uch changes are up to individual node operators as they require no deviation=
 from consensus. Yet ultimately this is a miner preference, and anyone can m=
ine. Finally, as I pointed out, burning is not necessary. Simply spending th=
e coin as a fee is sufficient.

> If an output is provably unspendable (burned) it is not a UTXO.
>=20
> It is worth noting that not all full node implementations require a store o=
f UTXOs, this is an implementation detail. For example, libbitcoin uses a fl=
ag on each output to indicate its spentness on the strong branch. As such th=
e store size is linear by height.
>=20
>> This will of course be remedied by compact UTXO representations later, bu=
t not today.
>> Similarly, it would be very nice to have non-0-amount `OP_RETURN` outputs=
, as `OP_RETURN` outputs are never stored in the UTXO database.
>> However, this will require a change in node relay policy, which again wil=
l take time to make possible, and would not be practical today.
>>=20
>> Thus I think use of UTXO is better than burning or mining-fee-spending.
>=20
> I don=E2=80=99t believe you have shown this.
>=20
> Best,
> Eric
>=20
>> Also, mostly trivia:
>> The use of UTXOs to advertise services is not original to me --- I found t=
he LN channel gossip to be the inspiration for this.
>> Publicly-announced channels indicate the backing UTXO that funds the chan=
nel.
>> The purpose of publicly announcing the channels is to be able to provide t=
he service, of forwarding across the Lightning Network; thus the public anno=
uncement serves as an advertisement for the service.
>> Channel closure immediately spends the UTXO, and also doubles to "revoke"=
 the existing "advertisement".
>> I found this ability to "revoke" the advertisement appealing, and thereby=
 designed the Bitcoin Classified Ads Network around the UTXO spentness mecha=
nism.
>>=20
>> Regards,
>> ZmnSCPxj