Mar 14, 2025 | RUMCG Meeting

Attendees

Nic Jansma, karlijn@rumvision.com, Andy Davies, Barry Pollard, Ben Zorb, Cliff Crocker, Dan Freeman, Eric Goldstein, Erwin Hoffman, Georgi Petrov, Gilberto Cocchi, Ivalo Hristov, Joan Leon, Mateusz Krzeszowiak, Magnus Dahl, Matt Zeunert, Rick Viscomi, Sergey Chernyshev, Sergio Gomes, Sia Karamalegos, Simeon Totev, Todd Gardner, Tony Conway, Yoav Weiss

Agenda

Baseline
Interop 2025
RUMCG issue tracking

Minutes

Admin

Welcomes all around, we have 26 people present
Agenda:

Baseline short update
Interop 2025
Positions, issues we want to track: what do we want to prioritize

Logistics

New repo: Github
Should we post to the repo in Github: benefit for easy to find for everyone, downside, more work. Looks like we are leaning to posting them there.

Next meeting: April 14 2025

What should we talk about next? Vote in the backlog, add a thumb up for the backlog.
End of the call decided: we are going to walk through Github and prioritize and discuss from there

Baseline

Tony Conway presents:

Baseline is a way to easily discover which features are save to use today. Baseline newly available (in 4 browsers, less than 30 months) Baseline Widely Available (more than 30 months). It works for Chrome, Edge, Safari and Firefox
Baseline is already used in MDN, can I use, web.dev, and also in RUM tools, like RUM archive and RUMvision
Also in developer tools ESLints new CSS linter, web platform dx repositories (in Github). Link: https://web-platform-dx.github.io/web-features/web-features/

Yoav: 3 ways to use:

“ Can I use” in RL
Breakage: what would break features across the web, is Baseline doing that? Like error reporting? It would be a nice way to use this.
Reporting API: extending this into Baseline? Things that are deprecated, does that also include it. Tony will come back to its Erwin Hofman mentions that it is included in the data.

Can you target users based on their Baseline version?
Server side serving the bundle that their Baseline would support, Sergio Gomes is thinking more Client side. Like image manage, serving different image formats
Yoav: https://github.com/whatwg/html/issues/4432
Sergey has the idea to bundle the features per categories. Progressive enhancements, but also RUM or performance, other W3C groups

Interop 2025

Support for LCP API and Event Timing API (For INP)
Barry said: Safari unknown timed (might be September), Firefox already has LCP.
Mattheus wonders if performance awareness/ eagerness to adopt performance will actually take a hit if Safari (which probably has good performance) becomes available.
CrUX will not show Safari data though (odds are very low that will ever be added, Barry says)
RUMarchive might give a good indicator of general worldwide CWV though
Yoav states: there are known discrepancies between safari and chrome in measurements, we should be aware of that
Yoav thinks that performance will still be bad in Safari in cases as well. Cliff also feels like it will give a more complete image, and is better for the field.
No Chrome on IOS yet. Chrome on IOS will not report to CrUX.
RUMCG issue tracking

Can we put out a position on CLS? Karlijn would like to, it’s missing in Safari statements. Seems like there is a consensus we should do this. Extra info:
https://github.com/whatwg/html/issues/4432
https://github.com/mozilla/standards-positions/issues/386

Karlijn: Info on when FF/Safari are going to implement this?
Barry: FF supports LCP, and ET but not interactionId
... Safari I suspect will be a big release, Sep this year or next
Sergey: Each vendor publishes their position on Interop
Mateusz: Existential concern about Safari implementing, will it increase performance (with faster devices), care less about CrUX?
... We have RUM in Admin panel, concern people will start caring less about performance in general when they see Safari/iOS traffic with fast devices
... We may have to double our efforts, for users using slower devices that are still important
... Couldn't see them before and now you can
Barry: Zero data Safari goes into CrUX
.. 0.01% FF data goes into CrUX
... RUM Archive can show differences between browsers
... Fact that CrUX won't include it, any SEO benefit will be in Chrome stats
... To Sergei's point, vendors don't make clear of things, other than we're supportive of these ones that made it through
... Published blog post in each, highlighted things
Yoav: 100% with Barry, but on top of that, right now because of what we discussed in WebPerfWG yesterday, there's some discrepancy around timeOrigin
... Safari starts timeOrigin after all Cross-Origin redirects are done, where Chrome and Firefox do not
... Could be a source of Safari being faster
... We need to be careful with reactions on one hand, if 50% of users have 0 performance problems it's great and we should celebrate
... I suspect it's not the case
... Given fact that no one has looked into Safari performance, there are things to fix
Mateusz: I wish it will be that way
... Collecting for more exotic browsers
... When Safari launches these metrics it may be the other way
Andy: I see Safari's support as an opportunity
Cliff: I think CrUX has been amazing, but it's been only half the picture
... I think more initiatives like RUM Archive can be a source of truth
... Helps balance the field a bit, info around other sources
... Lean more heavily into things like RUM Archive and data sharing
Matt: Core Web Vitals reporting on iOS in general?
... My understanding in EU Apple is allowing other browser engines, still a wrapper around Safari engine
Barry: There isn't a Chrome on iOS, unless attitudes change
... That'll be an interesting thing, once this does launch Chrome on iOS, will report Core Web Vitals even though it's not Chrome
... Don't know if they can or will support it in Europe
Yoav: Chromium on iOS prototype exists
Sergey: Short mention of topics discussed in adjacent WebPerfWG
Karlijn: Why wasn't CLS picked up?
Barry: Much more complicated to pick up in other browsers
... WebPerf signals for that metric less useful
Nic: Not sure if we'd make progress
Yoav: One thing is we could file for a standards position on this
... Nothing for LayoutShift
... Firefox in past were concerned about performance
... Chrome team showed how performance wasn't as affected
... Low on priorities list
... Filing standards positions is a good first step
Barry: Repo may not have existed
Yoav: Issue on Moz's repo but not on WebKit
Sergey: LCP and INP are load and interactivity have more vendors alignment while CLS has less parties expressing that interest, feels more designer backed, or UX backed
Nic: Harder to quantify frustration

RUMCG issue tracking

Cliff: Centralize issues that this group cares about
... Bugs that Andy is filing every day
... Other areas of interest for us
... Collective approach of taking ideas to WebPerfWG or vendors that are interested
... Could we push to fund these things?
... We've had lists and presentations
... Can kind of see that there's not a collective place
... We have a repo in Github Projects
Also adding this to the REPO. Talking about the Repo, the chairs propose using Github, where we have our own repo, to keep track of our positions/ Pain points:
Anyone can add issues. It’s also nice that we can track other repos. Looks like everyone is enthusiastic. We are going to be workin on it with the next session, the chairs will do some homework with setting up an ingest. This will also be our main focus for next months agenda.
Thanks all around, see you next month!