[bitcoin-dev] BIP proposal for Lightning-oriented multiaccount multisig HD wallets

[bitcoin-dev] BIP proposal for Lightning-oriented multiaccount multisig HD wallets

[Simone Bronzini]

[-- Attachment #1: Type: text/plain, Size: 3490 bytes --]


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi all,
last month we started looking for feedback (here and on other channels)
about a proposal for a new structure to facilitate the management of
different multisig accounts under the same master key, avoiding key
reuse but still allowing cosigners to independently generate new
addresses. While previously multiaccount multisig wallets were little
used, now that LN is becoming a reality it is extremely important to
have a better multiaccount management method to handle multiple payment
channels.
Please have a look at the draft of the BIP at the link below:

https://github.com/chainside/BIP-proposal/blob/master/BIP.mediawiki

Any feedback is highly appreciated, but in particular we would like to
collect opinions about the following issues:

1. coin_type level:
this level is intended to allow users to manage multiple
cryptocurrencies or forks of Bitcoin using the same masterkey (similarly
to BIP44). We have already received some legit objections that, since we
are talking about a Bitcoin Improvement Proposal,  it shouldn't care
about alt-coins. While we can agree with such objections, we also
believe that having a coin_type level improves interoperability with
muti-currency wallets (which is good), without any major drawback.
Moreover, even a Bitcoin maximalist may hold multiple coins for whatever
reason (short term speculation, testing, etc).

2. SegWit addresses:
since mixing SegWit and non-SegWit addresses on the same BIP44 structure
could lead to UTXOs not being completely recognised by old wallets,
BIP49 was proposed to separate the key space. Since this is a new
proposal, we can assume that wallets implementing it would be
SegWit-compatible and so there should be no need to differetiate between
SegWit and non-SegWit pubkeys. Anyway, if someone believes this problem
still holds, we thought about two possible solutions:
    a. Create separate purposes for SegWit and non SegWit addresses
(this would keep the same standard as BIP44 and BIP49)
    b. Create a new level on this proposed structure to divide SegWit
and non SegWit addresses: we would suggest to add this new level between
cosigner_index and change

We believe solution b. would be better as it would give the option of
having a multisig wallet with non SegWit-aware cosigners without having
to use two different subtrees.

This proposal is a work in progess so we would like to receive some
feedback before moving on with proposing it as a BIP draft.

Simone Bronzini

-----BEGIN PGP SIGNATURE-----
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
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=XgcB
-----END PGP SIGNATURE-----


[-- Attachment #2: 0xB2E60C73.asc --]
[-- Type: application/pgp-keys, Size: 15783 bytes --]

[-- Attachment #3: 0xB2E60C73.asc.sig --]
[-- Type: application/pgp-signature, Size: 566 bytes --]

Re: [bitcoin-dev] BIP proposal for Lightning-oriented multiaccount multisig HD wallets

[Luke Dashjr]

> Status: Proposed

This should only be set after peer review and implementations are complete, 
and you intend that there will be no further changes.

> As registered coin types we propose the ones already used for BIP44, which 
can be found at the following page.

I suggest just referring to SLIP 44 directly.

You're missing the Backward Compatibility and Copyright sections.



On Tuesday 29 August 2017 10:19:10 AM Simone Bronzini via bitcoin-dev wrote:
> Hi all,
> last month we started looking for feedback (here and on other channels)
> about a proposal for a new structure to facilitate the management of
> different multisig accounts under the same master key, avoiding key
> reuse but still allowing cosigners to independently generate new
> addresses. While previously multiaccount multisig wallets were little
> used, now that LN is becoming a reality it is extremely important to
> have a better multiaccount management method to handle multiple payment
> channels.
> Please have a look at the draft of the BIP at the link below:
> 
> https://github.com/chainside/BIP-proposal/blob/master/BIP.mediawiki
> 
> Any feedback is highly appreciated, but in particular we would like to
> collect opinions about the following issues:
> 
> 1. coin_type level:
> this level is intended to allow users to manage multiple
> cryptocurrencies or forks of Bitcoin using the same masterkey (similarly
> to BIP44). We have already received some legit objections that, since we
> are talking about a Bitcoin Improvement Proposal,  it shouldn't care
> about alt-coins. While we can agree with such objections, we also
> believe that having a coin_type level improves interoperability with
> muti-currency wallets (which is good), without any major drawback.
> Moreover, even a Bitcoin maximalist may hold multiple coins for whatever
> reason (short term speculation, testing, etc).
> 
> 2. SegWit addresses:
> since mixing SegWit and non-SegWit addresses on the same BIP44 structure
> could lead to UTXOs not being completely recognised by old wallets,
> BIP49 was proposed to separate the key space. Since this is a new
> proposal, we can assume that wallets implementing it would be
> SegWit-compatible and so there should be no need to differetiate between
> SegWit and non-SegWit pubkeys. Anyway, if someone believes this problem
> still holds, we thought about two possible solutions:
>     a. Create separate purposes for SegWit and non SegWit addresses
> (this would keep the same standard as BIP44 and BIP49)
>     b. Create a new level on this proposed structure to divide SegWit
> and non SegWit addresses: we would suggest to add this new level between
> cosigner_index and change
> 
> We believe solution b. would be better as it would give the option of
> having a multisig wallet with non SegWit-aware cosigners without having
> to use two different subtrees.
> 
> This proposal is a work in progess so we would like to receive some
> feedback before moving on with proposing it as a BIP draft.
> 
> Simone Bronzini

Re: [bitcoin-dev] BIP proposal for Lightning-oriented multiaccount multisig HD wallets

[Thomas Voegtlin]

On 29.08.2017 12:19, Simone Bronzini via bitcoin-dev wrote:

> 2. SegWit addresses:
> since mixing SegWit and non-SegWit addresses on the same BIP44 structure
> could lead to UTXOs not being completely recognised by old wallets,
> BIP49 was proposed to separate the key space. 

This will lead to old UTXOs not being recognized by NEW wallets, because
at some point new wallets will not care about implementing old standards.

The only way to address this is to get out of bip39 and bip43, and to
include a version number in the mnemonic seed.

Re: [bitcoin-dev] BIP proposal for Lightning-oriented multiaccount multisig HD wallets

[Simone Bronzini]

[-- Attachment #1.1.1: Type: text/plain, Size: 3575 bytes --]

Thanks for your feedback, I fixed what you suggested. As for the purpose
how should we move on? We would be inclined to use 46, but of course we
are open to any other number.


On 29/08/17 22:07, Luke Dashjr via bitcoin-dev wrote:
>> Status: Proposed
> This should only be set after peer review and implementations are complete, 
> and you intend that there will be no further changes.
>
>> As registered coin types we propose the ones already used for BIP44, which 
> can be found at the following page.
>
> I suggest just referring to SLIP 44 directly.
>
> You're missing the Backward Compatibility and Copyright sections.
>
>
>
> On Tuesday 29 August 2017 10:19:10 AM Simone Bronzini via bitcoin-dev wrote:
>> Hi all,
>> last month we started looking for feedback (here and on other channels)
>> about a proposal for a new structure to facilitate the management of
>> different multisig accounts under the same master key, avoiding key
>> reuse but still allowing cosigners to independently generate new
>> addresses. While previously multiaccount multisig wallets were little
>> used, now that LN is becoming a reality it is extremely important to
>> have a better multiaccount management method to handle multiple payment
>> channels.
>> Please have a look at the draft of the BIP at the link below:
>>
>> https://github.com/chainside/BIP-proposal/blob/master/BIP.mediawiki
>>
>> Any feedback is highly appreciated, but in particular we would like to
>> collect opinions about the following issues:
>>
>> 1. coin_type level:
>> this level is intended to allow users to manage multiple
>> cryptocurrencies or forks of Bitcoin using the same masterkey (similarly
>> to BIP44). We have already received some legit objections that, since we
>> are talking about a Bitcoin Improvement Proposal,  it shouldn't care
>> about alt-coins. While we can agree with such objections, we also
>> believe that having a coin_type level improves interoperability with
>> muti-currency wallets (which is good), without any major drawback.
>> Moreover, even a Bitcoin maximalist may hold multiple coins for whatever
>> reason (short term speculation, testing, etc).
>>
>> 2. SegWit addresses:
>> since mixing SegWit and non-SegWit addresses on the same BIP44 structure
>> could lead to UTXOs not being completely recognised by old wallets,
>> BIP49 was proposed to separate the key space. Since this is a new
>> proposal, we can assume that wallets implementing it would be
>> SegWit-compatible and so there should be no need to differetiate between
>> SegWit and non-SegWit pubkeys. Anyway, if someone believes this problem
>> still holds, we thought about two possible solutions:
>>     a. Create separate purposes for SegWit and non SegWit addresses
>> (this would keep the same standard as BIP44 and BIP49)
>>     b. Create a new level on this proposed structure to divide SegWit
>> and non SegWit addresses: we would suggest to add this new level between
>> cosigner_index and change
>>
>> We believe solution b. would be better as it would give the option of
>> having a multisig wallet with non SegWit-aware cosigners without having
>> to use two different subtrees.
>>
>> This proposal is a work in progess so we would like to receive some
>> feedback before moving on with proposing it as a BIP draft.
>>
>> Simone Bronzini
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev


[-- Attachment #1.1.2: 0xB2E60C73.asc --]
[-- Type: application/pgp-keys, Size: 15783 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 898 bytes --]

Re: [bitcoin-dev] BIP proposal for Lightning-oriented multiaccount multisig HD wallets

[Simone Bronzini]

[-- Attachment #1.1.1: Type: text/plain, Size: 1375 bytes --]

> This will lead to old UTXOs not being recognized by NEW wallets, because
> at some point new wallets will not care about implementing old standards.

Your observations make perfect sense. That's exactly why we endorse
option b. in my previous email.

> The only way to address this is to get out of bip39 and bip43, and to
> include a version number in the mnemonic seed.

As for the idea of having a versioning on mnemonic seeds, I believe it
would be a very useful feature indeed. How about opening a new,
separate, topic about it?


On 30/08/17 12:07, Thomas Voegtlin via bitcoin-dev wrote:
>
> On 29.08.2017 12:19, Simone Bronzini via bitcoin-dev wrote:
>
>> 2. SegWit addresses:
>> since mixing SegWit and non-SegWit addresses on the same BIP44 structure
>> could lead to UTXOs not being completely recognised by old wallets,
>> BIP49 was proposed to separate the key space. 
> This will lead to old UTXOs not being recognized by NEW wallets, because
> at some point new wallets will not care about implementing old standards.
>
> The only way to address this is to get out of bip39 and bip43, and to
> include a version number in the mnemonic seed.
>
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev


[-- Attachment #1.1.2: 0xB2E60C73.asc --]
[-- Type: application/pgp-keys, Size: 15783 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 898 bytes --]