Tim Dresser

Come work with us!

@anniesullie @igrigorik @slightlylate @yoavweiss @atannerhodges @jaffathecake Definition: IMO Performance > Latency, e.g., power, memory. Purpose: Access is probably most important, but broad UX improvements matter too. Focus on the tail is often effective, but chasing too far into the tail reduces actionability. Got to find the sweet spot!

@igrigorik @slightlylate @yoavweiss @atannerhodges @jaffathecake @tdresser This is a much clearer version of what I was trying to say earlier :)

👏 Thank you, @slightlylate! These guiding ideas are crucial. Excited to continue building on them—especially the principles & outcomes. 🔗 A Unified Theory of Web Performance calendar.perfplanet.com/2021/a-unified…

Want to improve Web UX and performance through Data Science and Software Development? Join Chrome’s Web Platform Speed Tooling team in Waterloo, Canada to gather data, prevent regressions and derive insight. Apply here, and email chromejobs@google.com: careers.google.com/jobs/results/1…

@digitarald @FirefoxDevTools I'm really sorry to hear that... I hope you're able to stay involved in the web.

Just jumped in the #MozillaLifeboat, after 9yrs open web. 3yrs @FirefoxDevTools Product Mgmt has been a 🚀. Guiding a passionate team & community to build a better web for all – all while honing many skills like research, analytics, outreach, growth, etc. 👋🏻 DM / hi@harald.me

📢Introducing Web Vitals, an initiative by Google to provide unified guidance for quality signals that are essential for great user experiences on the web. More at blog.chromium.org/2020/05/introd…

Chrome Speed Metrics has a new website at chromium.googlesource.com/chromium/src/+…. It includes links to webperf objectives for the quarter and a metrics changelog! Check it out and let me know what if other useful info could be added there.

Fruits of the last couple of months work for @Wikimedia: Paint Timing spec w3c.github.io/paint-timing/ is much more mature, and landed in @webkit trac.webkit.org/changeset/2608…. If all goes well one of the next Safari Tech Previews would have paint timing as an experimental feature. Yay!

@slightlylate @matthewcp @yoavweiss @AmeliasBrain @fantasai @n8Schloss Absolutely happy to chat about alternative solutions to this set of problems. You're also welcome to participate in the web performance working group calls, where most of the collaboration around this has taken place. (docs.google.com/document/d/10d…).

Open Standards aren't about posting a publicly-readable spec. They're about getting critiques and and consensus from an open community of stakeholders (including multiple implementers).

We worked closely with engineers at Chrome to develop the origin trial for isInputPending, a new API that addresses queueing time delays. code.fb.com/developer-tool…

@addyosmani @mikesherov @domenic @yoavweiss Thanks, I agree this is confusing. We'll get a note into the spec at the very least. (And I suspect Addy can make sure we update something more developer facing.)

@mikesherov @domenic cc @tdresser @yoavweiss

Seems tracking first-contentful-paint is a bit less straightforward then I had imagined... github.com/w3c/paint-timi… /cc @domenic

If you're interested in more detail on Chrome 69's FCP issues (developers.google.com/web/updates/20…), take a look at the postmortem (docs.google.com/document/d/1D3…)

Chromium intent to implement ElementTiming. I'm _very_ excited about this! Every pixel doesn't have the same value - now we'll be able to track hero images with RUM: groups.google.com/a/chromium.org…

🥳 The 201806 dataset is live and it has a brand new metric: First Input Delay! Read all about the new metric and how to explore it in the Chrome UX Report here developers.google.com/web/updates/20…

The event timing (github.com/wicg/event-tim…) origin trial is open for signup! Gain insight into your web app's event latency and First Input Delay. Sign up at bit.ly/event-timing-ot.

A cool visualization of HTTP Archive data about the usage trend of different libraries on the top sites. tdresser.github.io/js-library-usa…

@igrigorik @yoavweiss @addyosmani @jaffathecake @markusstaab For local testing, I'd support adding something to the devtools protocol or similar. We have a lot more flexibility regarding security and privacy if we don't ship to the wild.

@yoavweiss @addyosmani @jaffathecake @markusstaab @tdresser right, reflow *could* be one of the contributors to a long task, but that's not the right API to surface a count.. if that's what we're after. re, Feature Policy: see github.com/WICG/feature-p….

@addyosmani @jaffathecake do browser support a js api to check how many reflows happened? Would be great thing to have to prevent regressing in this area / maybe also usefull for unit testing..

@yoavweiss @addyosmani @jaffathecake @markusstaab @igrigorik Chrome's implementation will surface long tasks in this case, but the spec needs some work to include this. We're working on the spec tweaks!

@addyosmani @jaffathecake @markusstaab @igrigorik Theoretically, LongTasks (w3c.github.io/longtasks/) can and should expose that, assuming the reflows trigger long tasks. Not sure if that's implemented in v1 though /cc @tdresser

Browser cake is the best kind of cake. twitter.com/madhava/status…

99'th Percentile First Contentful Paint via Chrome User Experience report. Query: goo.gl/JE8UDD , Chart: goo.gl/qcKu1o.

record real-world perf traces via {Resource,Paint,Navigation,User}Timing: bit.ly/2sZxhyY - awesome work by @tdresser & @paul_irish!