From: Luke Dashjr <luke@dashjr.org>
To: Justus Ranvier <justus@openbitcoinprivacyproject.org>
Date: Thu, 22 Oct 2015 20:43:16 +0000
User-Agent: KMail/1.13.7 (Linux/4.1.9-gentoo-r1; KDE/4.14.8; x86_64; ; )
References: <201510220554.00367.luke@dashjr.org>
	<5628F8D2.1010709@openbitcoinprivacyproject.org>
In-Reply-To: <5628F8D2.1010709@openbitcoinprivacyproject.org>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-15"
Content-Transfer-Encoding: 7bit
Message-Id: <201510222043.17582.luke@dashjr.org>
Cc: Bitcoin Dev <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] [Bitcoin-development] Reusable payment codes
Precedence: list

On Thursday, October 22, 2015 2:55:14 PM Justus Ranvier wrote:
> On 22/10/15 00:53, Luke Dashjr wrote:
> > Sorry for the late review. I'm concerned with the "notification address"
> > requirement, which entails address reuse and blockchain spam. Since it
> > entails address reuse, the recipient is forced to either leave them
> > unspent forever (bloating the UTXO set), or spend it which potentially
> > compromises the private key, and (combined with the payment code)
> > possibly as much as the entire wallet.
> > 
> > Instead, I suggest making it a single zero-value OP_RETURN output with
> > two pushes: 1) a hash of the recipient's payment code, and 2) the
> > encrypted payment code. This can be searched with standard bloom
> > filters, or indexed with whatever other optimised algorithms are
> > desired. At the same time, it never uses any space in the UTXO set, and
> > never needs to be
> > spent/mixed/dusted.
> 
> The notification transaction portion is my least-favorite portion of the
> spec, but I don't see any alternatives that provide an unambiguous
> improvement, including your suggestion.
> 
> One of the most highly-weighted goals of this proposal is to be usable
> on as many mobile/light wallets as possible.
> 
> I know for sure that all existing platforms for balance querying index
> by address. Support for bloom filters or other querying methods is less
> comprehensive, meaning the set of wallets that can support payment codes
> would be smaller.

No, they just need to improve their software, and only to support receiving 
with payment codes (not sending to them). BIPs should in general not be 
designed around current software, especially in this case where there is no 
benefit to doing so (since it requires software upgrades anyway).

Luke