Re: CBOR-LD for VC

On Mon, Feb 14, 2022 at 2:18 AM Joel Thorstensson <oed@3box.io> wrote:

> Would be interesting to explore what the intersection of CBOR-LD, VCs, and
> IPLD would be. IPLD is a dataformat for hash linked data. It specifies a
> DAG-CBOR encoding which uses tag 42 to represent a hash link (which imo is
> pretty much the only way you can responsibly do linked data).
>

We decided not using the old version of CBOR-LD as it was a mishmash of
JSON-LD & IPLD, plus a little CBOR. Could be a newer version is better, but
we are very uncomfortable in supporting IPLD in the mashup.

We don’t have a replacement but are investigating a number of more CBOR
centric approaches, COSE, CSF (
https://test.webpki.org/csf-lab/home), or maybe one of our own tentatively
called crypto-envelope that supports both signatures & encryption, and
allows for partial signatures, multisig, transcryption, smart signatures
predicates, etc.

We also have some cryptographic objects defined in CBOR that this group
might find useful at
https://github.com/BlockchainCommons/Research/blob/master/papers/bcr-2020-006-urtypes.md
starting at 300.

These are largely for airgap QR wallet oriented scenarios: request a seed,
Shamir shard a seed, derive an hd-key, key, or address, join in a multisig
account and sign your part of one (though the later two are Bitcoin-centric
now they are designed to be adaptable in future), and animated QR for
larger transfers.

— Christopher Allen

>