[Bitcoin-development] Request For Discussion / BIP number - Multi-Currency Hierarchy For Use In Multisignature Deterministic Wallets

William Swanson swansontec at gmail.com
Wed Apr 8 16:41:00 UTC 2015


Oops, sorry I missed that.

Since that's the reason this proposal exists, I would consider putting
it right up top where people can see it. Also, since this proposal is
specifically designed for multi-sig, I would look at what BIP45 is
doing and maybe incorporate a "cosigner_index" branch. Otherwise, this
idea seems like a reasonable way to organize a wallet.

-William

On Wed, Apr 8, 2015 at 9:28 AM, 木ノ下じょな <kinoshitajona at gmail.com> wrote:
> William,
>
> I believe the reasoning for this is stated in the Coin Type section.
>
> "Public derivation is used so that cosigners need only know one of each
> other's public keys, rather than needing to distribute public keys for each
> coin."
>
> BIP44 has a coin level, but it's a private derived level, so cosigners would
> not be able to generate multiple crypto currencies of each others' without
> giving each other n xpubs where n is the number of currencies shared. This
> new proposal basically sticks coin type on the public derivation side of
> things so that I could generate litecoin or darkcoin multisigs without your
> permission...
>
> Kefkius,
>
> This BIP seems like a good fit for multi-currency wallets based on multisig.
> So kudos for putting it in writing.
>
> However, I don't know if this is really a BIP thing. It's not improving
> Bitcoin (Bitcoin Improvement Proposal... remember?), in fact, by definition
> it is improving altcoin usability.
>
> For that reason alone I will say I disagree for a BIP for this.
> - Jona
>
>
> 2015-04-08 16:46 GMT+09:00 William Swanson <swansontec at gmail.com>:
>>
>> It's not really clear why this is better than BIP 44 as it already
>> stands. You have the same fields, but they are just in a different
>> order. Couldn't you just use the existing BIP 44 hierarchy, but add
>> the convention that "wallet/account N" is the same wallet in each
>> supported currency?
>>
>> For example, if I have a wallet called "business expenses", which
>> happens to be wallet m / 44' / 0' / 5', for Bitcoin, then the same
>> wallet would be m / 44' / 3' / 5' for Dogecoin, and m / 44' / 2' / 5'
>> for Litecoin.
>>
>> I am trying to think of examples where your proposal is better than
>> BIP 44, but I can't think of any. Even backup recovery works fine. I
>> assume that your idea is to continue iterating over the different
>> wallet indices as long as you are finding funds in *any* currency.
>> Well, you can still do that with BIP 44. The fields are in a different
>> order, but that doesn't affect the algorithm in any way.
>>
>> Maybe you have some deeper insight I'm not seeing, but if so, you need
>> to clearly explain that in your motivation section. The current
>> explanation, "This limits the possible implementations of
>> multi-currency, multisignature wallets," is pretty vauge. Also, there
>> is nothing in this spec that addresses the multisignature use-case.
>> The BIP 45 spec does a lot of extra work to make multisignature work
>> smoothly.
>>
>> I'm not trying to criticize your proposal. I'm just trying to
>> understand what it's trying to accomplish.
>>
>> -William Swanson
>>
>>
>> On Wed, Apr 8, 2015 at 12:05 AM, Kefkius <kefkius at maza.club> wrote:
>> > I have a potential BIP, "Multi-Currency Hierarchy For Use In
>> > Multisignature Deterministic Wallets." I'm requesting discussion on it,
>> > and possibly assignment of a BIP number.
>> >
>> > It's located in this github gist:
>> > https://gist.github.com/Kefkius/1aa02945e532f8739023




More information about the bitcoin-dev mailing list