From: Luke Dashjr <luke@dashjr.org>
To: Johnson Lau <jl2012@xbt.hk>
Date: Tue, 19 Feb 2019 20:24:12 +0000
User-Agent: KMail/1.9.10 (enterprise35 0.20100827.1168748)
References: <9F8C0789-48E9-448A-A239-DB4AFB902A00@xbt.hk>
	<201902191904.04412.luke@dashjr.org>
	<B58087BD-50F2-40DF-BDF0-8D1FDFDCDE03@xbt.hk>
In-Reply-To: <B58087BD-50F2-40DF-BDF0-8D1FDFDCDE03@xbt.hk>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Message-Id: <201902192024.13243.luke@dashjr.org>
Cc: bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] Safer NOINPUT with output tagging
Precedence: list

Even besides NOINPUT, such a wallet would simply never show a second paymen=
t=20
to the same address (or at least never show it as confirmed, until=20
successfully spent).

At least if tx versions are used, it isn't possible to indicate this=20
requirement in current Bitcoin L1 addresses. scriptPubKey might not be=20
impossible to encode, but it isn't really clear what the purpose of doing s=
o=20
is.

If people don't want to use NOINPUT, they should just not use it. Trying to=
=20
implement a nanny in the protocol is inappropriate and limits what develope=
rs=20
can do who actually want the features.

Luke


On Tuesday 19 February 2019 19:22:07 Johnson Lau wrote:
> This only depends on the contract between the payer and payee. If the
> contract says address reuse is unacceptable, it=E2=80=99s unacceptable. I=
t has
> nothing to do with how the payee spends the coin. We can=E2=80=99t ban ad=
dress
> reuse at protocol level (unless we never prune the chain), so address reu=
se
> could only be prevented at social level.
>
> Using NOINPUT is also a very weak excuse: NOINPUT always commit to the
> value. If the payer reused an address but for different amount, the payee
> can=E2=80=99t claim the coin is lost due to previous NOINPUT use. A much =
stronger
> way is to publish the key after a coin is well confirmed.
>
> > On 20 Feb 2019, at 3:04 AM, Luke Dashjr <luke@dashjr.org> wrote:
> >
> > On Thursday 13 December 2018 12:32:44 Johnson Lau via bitcoin-dev wrote:
> >> While this seems fully compatible with eltoo, is there any other
> >> proposals require NOINPUT, and is adversely affected by either way of
> >> tagging?
> >
> > Yes, this seems to break the situation where a wallet wants to use
> > NOINPUT for everything, including normal L1 payments. For example, in t=
he
> > scenario where address reuse will be rejected/ignored by the recipient
> > unconditionally, and the payee is considered to have burned their
> > bitcoins by attempting it.
> >
> > Luke