To: Luke Dashjr <luke@dashjr.org>,
	Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
References: <e43c6e06-0bdc-360e-eb5c-a3726e4f0fc8@osc.co.cr>
	<201709292103.36630.luke@dashjr.org>
From: Dan Libby <dan@osc.co.cr>
Message-ID: <dcfaa57d-3fcd-973e-2548-5f9f318c0682@osc.co.cr>
Date: Fri, 29 Sep 2017 15:13:47 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
	Thunderbird/52.3.0
MIME-Version: 1.0
In-Reply-To: <201709292103.36630.luke@dashjr.org>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Subject: Re: [bitcoin-dev] Paper Wallet support in bitcoin-core
Precedence: list

It seems to me that the same statement can be made for *any* key storage
mechanism depending on one's security/threat model, including
bitcoin-core's internal wallet storage.  There certainly are cases where
a paper (or metal) offline wallet makes a lot of sense, particularly for
long-term offline storage... something that electronic media pretty much
sucks at.

Though if you care to elaborate I'd be interested to learn of your
specific critiques, if you have any beyond the generic statements here:
https://en.bitcoin.it/wiki/Paper_wallet

Regardless, the APIs I've proposed have uses beyond paper wallets.  It
can also be used by third party wallets, or any number of reasons that
individuals or devs might have to generate keys.


On 09/29/2017 02:03 PM, Luke Dashjr wrote:
> Paper wallets are a safety hazard, insecure, and generally not advisable.
> 
> 
> On Friday 29 September 2017 5:29:17 PM Dan Libby via bitcoin-dev wrote:
>> Hi,
>>
>> I'm writing to suggest and discuss the addition of paper wallet
>> functionality in bitcoin-core software, starting with a single new RPC
>> call: genExternalAddress [type].
>>
>> -- rationale --
>>
>> bitcoin-core is the most trusted and most secure bitcoin implementation.
>>
>> Yet today (unless I've missed something) paper wallet generation
>> requires use of third party software, or even a website such as
>> bitaddress.org.  This requires placing trust in an additional body of
>> code from a less-trusted and less peer-reviewed source.  Ideally, one
>> would personally audit this code for one's self, but in practice that
>> rarely happens.
>>
>> In the case of a website generator, the code must be audited again each
>> time it is downloaded.  I cannot in good faith recommend to anyone to
>> use such third party tools for wallet generation.
>>
>> I *would* recommend for others to trust a paper wallet that uses
>> address(es) generated by bitcoin-core itself.
>>
>> At least for me, this requirement to audit (or implicitly trust) a
>> secondary body of bitcoin code places an additional hurdle or
>> disincentive on the use of paper wallets, or indeed private keys
>> generated outside of bitcoin-core for any purpose.
>>
>> Unfortunately, one cannot simply use getnewaddress, getaccountaddress,
>> or getrawchangeaddress for this purpose, because the associated private
>> keys are added to the bitcoin-core wallet and cannot be removed... or in
>> the case of hd-wallets are deterministically derived.
>>
>> As such, I'm throwing out the following half-baked proposal as a
>> starting point for discussion:
>>
>>
>> -----
>>
>>     genexternaladdress ( "type" )
>>
>>     Returns a new Bitcoin address and private key for receiving
>>     payments. This key/address is intended for external usage such as
>>     paper wallets and will not be used by internal wallet nor written to
>>     disk.
>>
>>     Arguments:
>>     1. "type"        (string, optional) one of: p2pkh, p2sh-p2wpkh
>>                                         default: p2sh-p2wpkh
>>
>>     Result:
>>     {
>>         "privKey"    (string) The private key in wif format.
>>         "address"    (string) The address in p2pkh or p2sh-p2wpkh
>>                               format.
>>     }
>>
>>     Examples:
>>     > bitcoin-cli genexternaladdress
>>
>> ----
>>
>> This API is simple to implement and use.  It provides enough
>> functionality for any moderately skilled developer to create their own
>> paper wallet creation script using any scripting language, or even for
>> advanced users to perform using bitcoin-cli or debug console.
>>
>> If consensus here is in favor of including such an API, I will be happy
>> to take a crack at implementing it and submitting a pull request.
>>
>> If anyone has reasons why it is a BAD IDEA to include such an RPC call
>> in bitcoind, I'm curious to hear it.
>>
>> Also, I welcome suggestions for a better name, or maybe there could be
>> some improvements to the param(s), such as calling p2sh-p2wpkh "segwit"
>> instead.
>>
>>
>> ---- further work ----
>>
>>
>> Further steps could be taken in this direction, but are not necessary
>> for a useful first-step.  In particular:
>>
>> 1. an RPC call to generate an external HD wallet seed.
>> 2. an RPC call to generate N key/address pairs from a given seed.
>> 3. GUI functionality in bitcoin-qt to facilitate easy paper wallet
>> generation (and printing?) for end-users, complete with nice graphics,
>> qr codes, etc.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> bitcoin-dev mailing list
>> bitcoin-dev@lists.linuxfoundation.org
>> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev


-- 
Dan Libby

Open Source Consulting S.A.
Santa Ana, Costa Rica
http://osc.co.cr
phone: 011 506 2204 7018
Fax: 011 506 2223 7359