# Meeting Notes 2024-08-20

## Attendees
👉 Add your names here
omri gazitt
@xmlgrrl
Vladi Berger
Atul T
Mohit Saxena
Alex Olivier
Phillip Messerschmidt
George Fletcher
Elie Azerad
Gerry Gebel

## Agenda
- Follow-ups from last week
- new authzen backend for interop scenario deployed, test cases passing for all implementations
- Implementer's draft submission update
- Feedback from product engineering
- If we want a gRPC binding in addition to HTTPS, we will need to design it now, in a much more intentional way, because going from gRPC to REST is easy but not the other way around. BTW, Atul mentioned that many IETF WGs are now contending with this issue and wished they thought about it earlier.
- On the boxcarring spec, there is some feedback that the “fail on first fail” use-case (which we did not specify as of yet) would be much easier to implement if the evaluations map was an array (as it was in the beginning). I know we went back and forth on this, but there is now some implementers feedback that suggests we look at this again.

## Notes
- Omri showed some updates to the authzen backend - not a breaking change
- Implementer's draft material was resubmitted based on updates requested by Mark H
- Engineering feedback
- Discussion on value of map vs. array
- AT: since we don't require JWS or JWE, could do AuthZEN via protobuf
- Issue: IDs can be polymorphic
- Issue: Subject and action could be more precise
- There was rough consensus that we should take the gRPC / protobuf issue seriously.
- Atul: have the co-chairs arrive at a consensus on whether we want to delay the spec to incorporate these changes
- George: it's important to think about this even at the cost of slightly delaying the 1.0 release, so it can be sustainable in the market
- Alex: reference for how we do it in Cerbos: https://github.com/cerbos/cerbos/blob/main/api/public/cerbos/engine/v1/engine.proto#L223

## Actions
- Co-chairs to form a point of view on whether we should refactor the current spec to make a future gRPC binding possible
- Next week's meeting may be canceled, depending on whether David is able to lead it (Omri and Gerry are both out)
- We will pick up the gRPC conversation on September 3