Intent to Experiment: Element Capture

Hello Blink owners,

We are asking for a breaking period of 3 weeks to this API, followed by a renewed experiment for the traditional 6 milestones (2024-May-14 to 2024-Jun-04). A request that seems similar to me was granted in January 2022 for another API, and later shipped successfully with consensus with Mozilla and Apple.

At the time of writing, we have public support for this API from such companies as Zoom, Jitsi, Mux, DialPad, Whereby, Intel and Tella. Tango’s co-founder wrote “I can't emphasize enough how instrumental this specification would be for our product and user experience.” However, none of them have signed up for the OT as of yet. 

From the OT perspective:

This API allows Web developers to build novel new features; however, it requires a non-trivial investment. We are hoping that after giving more time, we will see more participation.

From the standardization perspective:

We need time to pick up discussions with Mozilla and Apple again. As Chrome’s Security and Privacy experts do not share the concerns Mozilla and Apple have previously voiced, it stands to reason that additional discussions will allow us to converge - and we will prioritize these discussions now.

Additionally, some possibilities remain for API changes that could perhaps allow for a compromise, mostly around cross-origin isolation. (Full disclosure - this is not my ideal outcome, but it’s a fallback possibility worth mentioning.)

Progress made:

• Spec: The spec has evolved and is now more mature, dealing better with such edge cases as loss of “eligibility for restriction.”

• TAG: Previously we held off on the request for a TAG review until we got some more developer feedback about the API shape. Having received this initial feedback, the TAG request has now been submitted.

• Signals: Signals have been requested. Mozilla responded. We intend to prioritize this discussion with them now.

• Feedback: Outreach for feedback from the spec community - multiple issues were filed on the spec by Web developers.

• WPT: Coverage has recently been extended.

Reasons to run a new trial:

Gain additional feedback from new participants. Examples for remaining areas where new feedback could help include:

• Uncover new edge cases which were not uncovered by the spec authors and reviewers, implementers and current OT participants. The current edge cases here demonstrate how non-obvious these may be.

• Validate (or refute) the assumptions we have made about the viability of an MVP that is missing some functionality. Examples:

  • Events notifying apps when an element stops/starts being “eligible for restriction”.

  • Mechanism to force elements into an “eligible for restriction” state.

• Encourage future OT participation by Web developers, by demonstrating that risks associated with relying on an origin trial, while real, are partially offset by a commitment to keep the OT going until discussions conclude.

Thanks,

Elad