Normative references, and stability criteria

[svgeesus]

From how to go to Proposed Recommendation:

In general, documents do not advance to Proposed Recommendation with normative references to other specifications that are considered unstable. See also Normative References Guidelines. Documents must not include normative references to a Rescinded/Obsolete/Superseded Recommendation.

In the CR snapshot:

• WebRTC 1.0 was a CR, is now a Rec. Bikeshed should pick up that change automatically.
• Media Capture and Streams is still a CR so we need to argue that it meets the stability criteria
• WebIDL is still a draft but that didn't stop WebRTC referencing it in their Rec so there is evidence it already meets the stability criteria.

[svgeesus]

I contacted the staff contacts for the Web Real-Time Communications Working Group asking about stability for the 28 Jan 2021 CR of Media Capture and Streams.

[rtoy]

Are those the only ones you found?

I'm also curious why WebIDL isn't automatically allowed.

[svgeesus]

Yes, those are the only ones I found.

WebIDL is in practice automatically allowed, I checked with some recent specifications that reference it.

[rtoy]

The media capture and streams spec appears to be referenced only from the privacy section. (Searching for "media capture" shows only 3 instances.) If so, we're not using media capture in any way within the spec, so perhaps it's ok that it's not a recommendation.

That leaves only WebIDL. (The spec already links to the WebRTC recommendation).

I think we're fine here? Is there more to do? If not can we close this?

[svgeesus]

We can close this.