[bitcoin-dev] A proposal for a "PSBT for descriptors" format

[bitcoin-dev] A proposal for a "PSBT for descriptors" format

[SeedHammer Team]

Hi,

At SeedHammer we're interested in standard, compact output descriptors to make
self-contained metal engraved backups feasible. To that end, we're proposing a
descriptor format based on the PSBT binary encoding. The format has not reached
widespread consensus, never mind adoption, so at this point we're soliciting
comments before formally proposing it as a BIP.

See [proposal], [implementation] and [playground] for details and examples.

The format is a binary and compact serialization specification for the
[wallet-policies] BIP. Features:

- Based on the binary [BIP174] PSBT format, including re-using the compact
  PSBT_GLOBAL_XPUB encoding for extended keys.
- The descriptor itself is encoded in the same textual format as described
  in BIPs 380-386 (+389).
- Key references (always) use the wallet-policies format @<key-index>.
- Miniscript is trivially supported, except inline keys are not allowed, and
  pk(NAME) expressions are replaced with indexed (@<idx>) key references.
- Metadata such as labels and birthdate blocks are encoded as PSBT
  map entries.

Known issues:

- CBOR vs PSBT. Blockchain Commons believes[0] that a CBOR based format is better
  because it is a widely used binary encoding standard, whereas we believe the
  complexity of CBOR doesn't justify its cost compared to the PSBT encoding
  already widely supported by wallet software.
- The proposal specifies a separate header and magic; should the format instead be
  an extension to the PSBT format?

Thanks,
E

[proposal] https://github.com/BlockchainCommons/Research/issues/135
[implementation] https://github.com/seedhammer/bip-serialized-descriptors
[playground] https://go.dev/play/p/nouZlbbcEWt
[wallet-policies] https://github.com/bitcoin/bips/blob/bb98f8017a883262e03127ab718514abf4a5e5f9/bip-wallet-policies.mediawiki
[BIP174] https://github.com/bitcoin/bips/blob/master/bip-0174.mediawiki

[0] https://github.com/BlockchainCommons/Research/issues/135#issuecomment-1789644032

Re: [bitcoin-dev] A proposal for a "PSBT for descriptors" format

[Brandon Black]

Hi Seedhammer,

I think the goal of such a format should be that it is already a valid
PSBT, or can be trivially converted into one. Ideally, a coordinating
device can load the standardized descriptor file, add inputs (PSBTv2) or
unsigned TX (PSBTv1), and send it to compatible signing devices without
further modification.

Seems like the additions to BIP174 would be a PSBT_GLOBAL_DESCRIPTOR
with key of <birthblock>|<name> and value of the descriptor encoded as
proposed, and then PSBT_GLOBAL_KEY_NAME with key of <fingerprint>
and value of <name>.

Best,

--Brandon

On 2023-11-23 (Thu) at 22:25:43 +0000, SeedHammer Team via bitcoin-dev wrote:
> Hi,
> 
> At SeedHammer we're interested in standard, compact output descriptors to make
> self-contained metal engraved backups feasible. To that end, we're proposing a
> descriptor format based on the PSBT binary encoding. The format has not reached
> widespread consensus, never mind adoption, so at this point we're soliciting
> comments before formally proposing it as a BIP.
> 
> See [proposal], [implementation] and [playground] for details and examples.
> 
> The format is a binary and compact serialization specification for the
> [wallet-policies] BIP. Features:
> 
> - Based on the binary [BIP174] PSBT format, including re-using the compact
>   PSBT_GLOBAL_XPUB encoding for extended keys.
> - The descriptor itself is encoded in the same textual format as described
>   in BIPs 380-386 (+389).
> - Key references (always) use the wallet-policies format @<key-index>.
> - Miniscript is trivially supported, except inline keys are not allowed, and
>   pk(NAME) expressions are replaced with indexed (@<idx>) key references.
> - Metadata such as labels and birthdate blocks are encoded as PSBT
>   map entries.
> 
> Known issues:
> 
> - CBOR vs PSBT. Blockchain Commons believes[0] that a CBOR based format is better
>   because it is a widely used binary encoding standard, whereas we believe the
>   complexity of CBOR doesn't justify its cost compared to the PSBT encoding
>   already widely supported by wallet software.
> - The proposal specifies a separate header and magic; should the format instead be
>   an extension to the PSBT format?
> 
> Thanks,
> E
> 
> [proposal] https://github.com/BlockchainCommons/Research/issues/135
> [implementation] https://github.com/seedhammer/bip-serialized-descriptors
> [playground] https://go.dev/play/p/nouZlbbcEWt
> [wallet-policies] https://github.com/bitcoin/bips/blob/bb98f8017a883262e03127ab718514abf4a5e5f9/bip-wallet-policies.mediawiki
> [BIP174] https://github.com/bitcoin/bips/blob/master/bip-0174.mediawiki
> 
> [0] https://github.com/BlockchainCommons/Research/issues/135#issuecomment-1789644032
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev