Re: [bitcoin-dev] New serialization/encoding format for key material

[Jonas Schnelli <dev@jonasschnelli.ch>]

[-- Attachment #1: Type: text/plain, Size: 2724 bytes --]

> I have some concerns about the use of Bech32. It is designed for
> detecting 3 errors up to length 1023 (but is then picked specifically
> to support 4 errors up to length 89). However, for error correction
> this translates to just being able to efficiently correct 1 error
> (3/2, rounded down) up to length 1023. You can of course always try
> all combinations of up to N changes to the input (for any N), testing
> the checksum, and comparing the results against the UTXO set or other
> wallet information that may have been recovered. However, the checksum
> at best gives you a small constant speedup here, not a fundamentally
> improved way for recovery.

Thanks Peter

I removed the part in the proposals that made false claims about the error
correction or cpu-intense key recovery.

I wrote some test code and figured out that my Core i7 machine can
do 31’775 operations per seconds of a addr-derivation-comparison
(bech32 decode, bip32 ckd, hash160, Base58check).
This is non-optimized code running non-parallelized.

Just in case someone wants to do more math here.

Without knowing to much about BCHs, ideally there would be a code that
includes the fact that computational costs for error correction can be very
high during a disaster recovery and that we can probably assume that the
user can provide a derivation element like a used address or pubkey.

Deriving one million child keys and comparing them against an address
table will take less than a minute on consumer systems.

> * correct 7 errors = 26 checksum characters (~ length * 1.25)
> 
> So it really boils down to a trade-off between length of the code, and
> recovery properties.

I think 5% error correction (7 errors at 555bits) with a 26 char checksum is
probably an acceptable tradeoff.

Resulting string with 26 checksum chars (mockup):
xp1qqqqqq8z4rsgv54z9a92yla4m2yrsqdlwdl7gn6qldvwkuh3zrg66z8ad2snf832tgaxcuv3kmwugzl5x8wtnkj2q3a03ky0kg8p7dvv4czpjqgvv4zgnvv4zgnvv4zgnvv4zgngn
(140 chars)

Versus the bech32 (6 char checksum):
xp1qqqqqq8z4rsgv54z9a92yla4m2yrsqdlwdl7gn6qldvwkuh3zrg66z8ad2snf832tgaxcuv3kmwugzl5x8wtnkj2q3a03ky0kg8p7dvv4czpjqgvv4zgn
(120 chars)

Versus an xpriv:
xprv9wHokC2KXdTSpEepFcu53hMDUHYfAtTaLEJEMyxBPAMf78hJg17WhL5FyeDUQH5KWmGjGgEb2j74gsZqgupWpPbZgP6uFmP8MYEy5BNbyET
(111 chars)

Not sure if the additional 20 characters make the UX worse.
Typing in +20 chars in a disaster recovery is probably acceptable.

> If there is interest, I can construct a code + implementation for any
> of these in a few days probably, once the requirements are clear.


Yes. Please.
Lets first wait for more feedback about the error robustness though.

Thanks
-
Jonas

[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]