2021-04-02 Paper Based Credentials Drafting Group Meeting Notes

Time	Item	Who
2 min	Welcome & Antitrust Policy Notice	Chair
5 min	Introductions	All
30 min	JSON Format Discussion (Review of David Janes "Thoughts & Notes" from Slack)	David Janes
20 min	Consensus Points (Review of David Janes "Thoughts & Notes" from Slack)	David Janes
3 min	Wrap Up	Chair

(PDFs posted)

1. Welcome and Linux Foundation antitrust policy

2. JSON Formats (Based on prompt)

JSON-LD enables a single representation of a credential that could work as both paper and digital (means trade-off on space, but also meant one single credential). Also enables zero knowledge proofs.
- You never present VC you present proof of the VC - important in that JSON-LD is common denominator and can help unify this across paper and digital
There will be multiple flavors in payloads and QR codes no matter what we do.
- Note that GHPs won’t necessarily have massive data elements (can be constrained if we focus on travel, not full use in epidemiology)
- QR codes serve two purposes: proof + also something you might use to load credential into wallet (and not every QR code needs to be translated back to digital)

3. Points to gain consensus on (based on prompt)

There’s never going to be a single set of schemas - we need to design a system that assumes a mess.
- Group should explore what can we do to address this issue, rather than assuming this needs to be standardized.
- Group should make a recommendation re: how to distinguish between payloads and codes.
Need to understand credential vs. pass

Pass produced not by the original source of a covid test but by some intermediary - a secondary issuer and a secondary credential that is typically used in a specific context for a specific purpose
Terminology can be confusing to customers (e.g., IBM had to walk this language back)

Pattern of recombining credentials into another credential will come up in a lot of discussion (passes derived from credentials)
- Need to better help people understand that verifiable credentials don’t go on paper, presentations do (will help better understand credentials vs. proofs)

5. Wrap up

David Janes Thoughts & Notes (Discussion Prompt)

Credentials transform into Paper and back again losslessly - e.g. to the GHPC defined W3C VC
Passes transform into Paper as a one-way operation (e.g. PathCheck)

If we have agreement on (2) and (2a) in particular, what are the ways of encoding the JSON

Ease of Implementation
Readily available and mature libraries in popular languages
Best Size of Compression
Size of Compress “Good Enough” (say, under 500 bytes - cross check with size GHPC is recommending!)
Amount of code required
Works in QR Code - Go/NoGo (cross check with size GHPC is recommending!)

Consensus Needed: is it OK if there are multiple compression methods - verifiers have a lot of work to do anyway?
Note that there are going to be multiple different QR payloads no matter what we do:

Consensus Needed: GPHC should make a recommendation how to distinguish between different payload types on QR codes, even if they are not using GHPC credentials.
PathCheck format:

Consensus Needed: is PathCheck format the recommended format for passes, or is it PathCheck + compressed W3CVC.
Question: is GHPC defining a pass payload, or just a credential?
Question: How does DIVOC fit in all this??
Future meetings, but not now: identity binding & for (2a) being able to self inspect payload for personal assurance

Action Items