Jon Johnson

> only the last one is considered the structured syntax This makes sense, the last one should be the format of the payload and trimming the suffix should yield the...

> Do you think the artifacts repo should provide guidance on the media types? Definitely, but they're going to look to image-spec for inspiration, and we should have a good...

Posting back here, @estesp kindly took some notes in the OCI call: https://hackmd.io/El8Dd2xrTlCaCG59ns5cwg#Notes1 Some action items from @dmcgowan 1. Add a description to image-spec of the media type format. We...

Something I just ran into re: nondistributable layers... this mostly involves distribution, but most of the relevant conversation has been happening in this repo. This PR says that unrecognized media...

> Do we care if OCI media types are valid RFC6838 media types? E.g. "zstd" isn't defined in the Structured Syntax Suffix Registry. FWIW [this is no longer true ](https://github.com/facebook/zstd/issues/1846)🎉

This is a real problem, not just for me: https://github.com/moby/buildkit/pull/1879 cc @tonistiigi come share in my pain

It seems like we _suggest_ this, already? From [here](https://github.com/opencontainers/image-spec/blob/master/config.md#properties): > * Labels object, OPTIONAL > > The field contains arbitrary metadata for the container. This property MUST use the [annotation...

@vbatts added this in https://github.com/opencontainers/image-spec/pull/371 -- do you have any opinions? I don't have much familiarity with how folks use labels in practice.

Regarding the harbor issue: just speaking from the perspective of a registry, I think that annotations are preferred here. Parsing labels requires reading a config blob, which is otherwise unnecessary...

The golang proposal has been accepted.