Date: Sun, 02 Aug 2020 14:22:30 +0000
To: Mike Brooks <m@ib.tc>
From: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Reply-To: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Message-ID: <-vfdrKtX7fPIncbPlX7CkHYD0NjZDz6VUjlkUrCO-feq3zaQpXRl8Wu_HP7CRXm1sc_p0B0IZjnTxiHfz_saXWL5W6ax9CVnddaDfPQvTBE=@protonmail.com>
In-Reply-To: <CALFqKjSrQ260XfYtdvd9N=ZKZvJqE2iz+reGtBNpq++br9S1ig@mail.gmail.com>
References: <CALFqKjTDcsD4xpc-TryTRCNnCJivyqkU9dv0ff2emP5kxZF1bA@mail.gmail.com>
 <1cxOISncF4dES_Ijm5FJUhwUAzBupnLPiv3qnU20as76zMMhVyggkg1hphq4ehuqEFK_H88TBbfI2pKbLWzzx7E6kOUXeC-yOcfrOkg3uAY=@protonmail.com>
 <CALFqKjSrQ260XfYtdvd9N=ZKZvJqE2iz+reGtBNpq++br9S1ig@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>
Subject: Re: [bitcoin-dev] Smaller Transactions with PubRef
Precedence: list

Good morning Mike,

The issue with SCRIPT re-evaluation is that reorgs cause more processing to=
 be done by nodes.

Floating-point Nakamoto Consensus does not help here, since a node can rece=
ive the lower-scored block first, and *then* a higher-scored block, and thu=
s will ***still*** observe a reorg since the chain tip is replaced with a h=
igher-scored block later.

This still increases the processing load on validating fullnodes, and preve=
nts any kind of pruning from working for validating fullnodes.

A miner can also still mount a DoS on validating fullnodes, with `OP_PUBREF=
` and Floating-Point Nakamoto Consensus by re-mining the same block, and br=
oadcasting a block if it has higher score than the previous chain tip.
This locks the blockchain ***and*** increases the load on fullnodes, which =
have to re-validate uses of `OP_PUBREF` that might refer to the chain tip.

Regards,
ZmnSCPxj

> Hey=C2=A0 ZmnSCPxj,
>
> Re-orgs should be solved in a different=C2=A0way.=C2=A0
>
> Best Regards,
> Micahel
>
> On Sat, Aug 1, 2020 at 5:36 PM ZmnSCPxj <ZmnSCPxj@protonmail.com> wrote:
>
> > Good morning Mike,
> >
> > Hard NAK.
> >
> > The responses to the original posting already pointed out important pro=
blems with this:
> >
> > * Encourages address reuse, hurting fungibility and privacy.
> > * Prevents pruning, since access to previous blocks must always be avai=
lable in order to validate.
> > * Optimized implementation requires creating yet another index to previ=
ous block data, increasing requirements on fullnodes.
> > * Requires SCRIPT to be re-evaluated on transactions arriving in=C2=
=A0 newblocks, to protect against reorgs of the chaintip, and in particular=
 `OP_PUBREF` references to near the chaintip.
> >
> > None of these issues have been addressed in your current proposal.
> > The proposal looks at clients only, without considering what validators=
 have to implement in order to validate new blocks with this opcode.
> >
> > Regards,
> > ZmnSCPxj