Agenda for Dec 2 meeting

foolip commented 2 years ago

Here's the proposed agenda/procedure for the Dec 2 meeting:

Show proposals and note which have "requires" labels
Discussion, enthusiasm or concerns for the remaining proposals from the previous meeting. If at all possible, please comment on the issue before the meeting and ensure they have the appropriate labels.
Indicating support for proposals.
- We take turns to enumerate which of the proposals we would actively like to see included in Interop 2022. "We" isn't limited to Apple, Google and Mozilla, anyone in a position to contribute to a major browser engine (Chromium, Gecko, WebKit) could chime in here.
Establishing the union of proposals with some support from the previous step
Does anyone have concerns about including any of the proposals now listed? Can those concerns be addressed, or is the feature itself the problem?
Establish which proposals are still on the table, and action items for next time, how to make a final decision.

Please comment on this issue with any changes/additions.

Previous meeting: https://github.com/web-platform-tests/interop-2022/issues/24

foolip commented 2 years ago

In order to do the "Indicating support for proposals" and "Does anyone have concerns" steps, it's important that everyone participating has made a list of proposals they want to see included and would rather not see included ahead of time. This is along the lines of what @jgraham suggested in the previous meeting.

meyerweb commented 2 years ago

@bkardell and I talked these over in depth, as well as the rest of the issues list outside of the “compat-bug-proposal”, and the following is what we came up with.

We’re strongly in favor of including all (or any) of the following:

dialog / ::backdrop
OffscreenCanvas
A Form Controls bucket
A typography bucket
A scrolling bucket
Improving the test infrastructure (ref.: pointer events)

That last one maybe isn’t an official part of Interop 2022, but we’d love to see it taken up, and be mentioned as part of planning for future Interops. I’ll talk about the “buckets” at the end of the post.

We’re also in favor of Cascade Layers, but with some reservations. It does have [three implementations, all at least somewhat skeletal](), so that’s a very strong reason to include it — there’s a foundation there on which advances could be quickly built. But also the spec feels like it’s flux, and we’re wary of including something that might run into spec-revision chop and halt implementation altogether, dragging down progress of the whole effort.

We’re iffy about the following, any of which we’d probably be happy to support if the cited drawbacks were resolved ahead of Interop 2022 launch:

AVIF support, because there have been vendor objections in the past and we haven’t heard of them being resolved.
Editing / execCommand, because while it would make life a lot better for a lot of users, and it has very nearly one hundred thousand tests in WPT, it apparently needs spec work. That brought us up short.
Color spaces / functions, because our understanding is that LAB support is a fundamental blocker on this, and LAB support is limited. If our understanding is correct, we’d be very much in favor of LAB support being on the Interop 2022 list. If it isn’t correct, let us know.
Box alignment on block boxes, due to being unclear on the state of testing.
Viewport measurement / units, because tests are needed and the scope of the problem feels like it lacks definition.

For varied reasons, we believe pointer events, subgrid, and container queries should be held for Interop 2023 (or later).

Now, to explain what we mean by the “buckets” in our first list. We think form controls, typography, and scrolling are all areas where advancement could do a lot of good. They’re all pain points for developers, and advancing in those areas would not just be a win for compatibility scores, but for the web in general. So we think those topic areas should be represented in Interop 2022. Whether that means one big issue for each, or a collection of known issues, is up for debate.

For example, we could decide for the “typography bucket” to include the font-variant-* properties, hanging-punctuation, and line clamping, while leaving out initial-letter and widows & orphans. For the “form controls bucket”, we could pick and choose from the quite frankly daunting list of possible things to fix. For scrolling, we could decide to take on all the scroll-related stuff, or just stick to one problem, like making scroll-snapping more consistent.

Hopefully that all made sense, but if anyone has any questions, let us know!

foolip commented 2 years ago

Thanks @meyerweb, that's very useful! I have thoughts and opinions about a few things, but I'll save that for Thursday after we see which proposals seem most viable.

foolip commented 2 years ago

Mostly in order to help me in the meeting, here's the list of proposals that Google is most interested in for Interop 2022:

web-platform-tests / interop

Agenda for Dec 2 meeting #36