Re: [bitcoin-dev] BIP49 Derivation scheme changes

[Dan Libby <dan@osc.co.cr>]

On 08/30/2017 12:24 AM, shiva sitamraju via bitcoin-dev wrote:

> What would happen if you recover a wallet  using seed words ?
>   1. Since there is no difference in seed words between segwit/non
> segwit, the wallet would discover both m/44' and m/49' accounts
>   2. Note that we cannot ask the user to choose an account he wants to
> operate on (Segwit/Non segwit). This is like asking him the HD
> derivation path and a really bad UI
>   3. The wallet now has to constantly monitor both m/44' and m/49'
> accounts for transactions

small nit with 3.

It seems to me that the wallet would perform initial discovery on m/44
and m/49, and then would find transactions at one or the other, so it
can then record the type somewhere and from then on need only monitor
one branch.

Still, I agree it is ugly, makes initial discovery up to 2x slower, etc.

> *- XPUB Derivation*
> This is something not addressed in the BIP yet.
> 
> 1. Right now you can get an xpub balance/transaction history. With m/49'
> there is no way to know whether an xpub is from m/44' or m/49'
> 
> 2. This breaks lots of things. Wallets like electrum/armory/mycelium
> <https://blog.trezor.io/using-mycelium-to-watch-your-trezor-accounts-a836dce0b954>support
> importing  xpub as a watch only wallet. Also services like
> blockonomics/blockchain.info <http://blockchain.info> use xpub for
> displaying balance/generating merchant addresses
> 
> Looking forward to hearing your thoughts

speaking as author of tools hd-wallet-addrs and hd-wallet-derive, I
agree this is problematic.

would be great if xpub/xprv could somehow encode their absolute path in
wallet for tools to read.  Users cannot be expected to know.