The performance team I lead at Automattic is hiring - automattic.com/work-with-us... #webperf #php #jobs #wordpress #woocommerce
03.03.2026 16:08 β π 5 π 5 π¬ 0 π 0The performance team I lead at Automattic is hiring - automattic.com/work-with-us... #webperf #php #jobs #wordpress #woocommerce
03.03.2026 16:08 β π 5 π 5 π¬ 0 π 0"Add a minimal privacy-preserving secure DNS telemetry to check current version" salsa.debian.org/php-team/php... - surprised and disappointed to see this #php #security #debian
26.02.2026 17:05 β π 0 π 0 π¬ 0 π 0Definitely feeling "the last 20% takes 80% of the work" right now.
24.02.2026 17:34 β π 0 π 0 π¬ 0 π 0
Many developers misinterpret what the DevTools performance test results tell them. My video below explains how Chrome implements network throttling and how it impacts your data.
Neither the bandwidth nor the latency implementation match what you'd see on a real slow network.
youtu.be/93HI4dnGjMk
Safari 26.3 is here with Zstandard file compression, Navigation APIβs AbortSignal for better control in single-page apps, and fixes for anchor positioning, multi-column layouts, & more. Plus fullscreen video now dims surroundings in visionOS.
webkit.org/blog/17798/webkit-features-for-safari-26-3/
It's been a while since my last blog post, so I made up for it with a wall of text π
This article looks at why JS-heavy web apps tend to miss their performance goals over time, what you can do to somewhat mitigate that, and ends with a plea for us to move back to the server, as an industry.
Container Timing is now officially Ready for Developer Testing in Chromium 145! π
We want your feedback on the design and usage. Check out the details and how to try it:
blogs.igalia.com/dape/2026/02...
cc @igalia.com @developer.chrome.com
Both of these websites have a 2-second LCP score β but the user experience is very different!
Learn about how to relate metrics to visitor experience on your website: www.debugbear.com/blog/web-per...
Why do sites keep offering to save or remember this computer, if they never, ever, ever actually remember this computer? So frustrating.
04.02.2026 23:52 β π 1 π 0 π¬ 1 π 0Days where it feels like LLMs might be the best rubber ducking experience so far - one that can talk back and often provide something useful.
26.01.2026 17:41 β π 2 π 0 π¬ 0 π 0That warm feeling of looking at a function and thinking to yourself - there is a very good chance I can make that faster.
20.01.2026 23:47 β π 3 π 0 π¬ 0 π 0Anyone has tried out Mago - mago.carthage.software - a linter/formatter/analyzer for PHP code? I gave it a very brief run against a single PHP file and was surprised at how aggressive the defaults are. It might be a very useful tool, but I'd need to config it to be more reasonable first. #php
15.01.2026 16:19 β π 1 π 0 π¬ 0 π 0The annual tradition of going back over my last 12 days of date references to fix the year in dates that I wrote down.
12.01.2026 15:17 β π 0 π 0 π¬ 0 π 0Compression Dictionaries with Zstandard in PHP josephscott.org/blog/2026/co... #webperf
07.01.2026 20:45 β π 1 π 0 π¬ 0 π 0I am starting to think that the iOS 26 OS/UI experience equivalent of the Apple butterfly keyboard and touch bar. iOS 27 could do a full UI revert to the iOS 18 UI and be heralded as tremendous improvement. #apple
15.12.2025 15:26 β π 1 π 0 π¬ 0 π 0How to Switch Debian 13 /tmp from tmpfs to disk - josephscott.org/blog/2025/ho...
05.12.2025 19:38 β π 2 π 0 π¬ 0 π 0With the #LogicMonitor ( couldn't find a Bluesky account ) acquisition of Catchpoint I have a request. Please revert the UI changes that were done to webpagetest.org recently. I just want my long trusted go to web perf tool back.
04.12.2025 23:23 β π 1 π 0 π¬ 0 π 0
I'm playing around with JPEG XL's "JPEG repackaging" feature. It repackaged a 147kb JPEG down to 126kb, but on my Pixel 10, the new file takes 78ms to decode, vs 10ms for the JPEG.
At 3g+ speeds⦠that feels like a performance loss.
random-stuff.jakearchibald.com/apps/img-dec...
So many times that my kids and I have made this joke π¦·ππ
03.12.2025 19:02 β π 3 π 0 π¬ 0 π 0I like the idea of `dns-persist-01` - datatracker.ietf.org/doc/html/dra... - "Unlike the existing "dns-01" challenge, this record is designed to persist and may be reused for multiple certificate issuances over an extended period." - one less moving part when generating a new cert.
02.12.2025 16:08 β π 2 π 0 π¬ 0 π 0
Inspired by perfnow.nl, I've dusted off drafts of my network and device situation analysis. Good news/bad news: devices and networks are improving, but pages are swelling. The web is usable for the wealthy, but less so for everyone else.
infrequently.org/2025/11/perf...
Out-of-order HTML patching, another proposal from TPAC, lets HTML be presented non-linearly, and a single file can update multiple parts of the page.
Would this feature be useful to you?
It can be truly painful to look at, easier yo avoid it and pretend it doesnβt exist
05.11.2025 17:20 β π 0 π 0 π¬ 0 π 0
#LCP and #INP support in #Safari really coming close now (maybe even in 2025?) as Safari 26.2 Beta Release Notes includes:
- support for the Event Timing API
- Largest Contentful Paint metric
Great work by the #Webkit team!
A blogpost:
www.rumvision.com/blog/safari-...
Realizing now that for my #webperf folks I should have dressed up in the scariest of all possible Halloween costume: your p99 metric value
05.11.2025 15:19 β π 1 π 0 π¬ 1 π 0Do future you ( and others ) a favor - check your return values. Funny how something that was an edge case that "never happens" at some point in the future starts happening every day.
03.11.2025 14:39 β π 0 π 0 π¬ 0 π 0This has been on my todo list for too long
24.10.2025 17:24 β π 0 π 0 π¬ 0 π 0Using the new WPT UI that @catchpointsystems.bsky.social just rolled out, and I'm angry and sad and frustrated and deeply confused all at once. What an utter debacle. Roll this back.
24.10.2025 16:47 β π 6 π 2 π¬ 5 π 0Tomorrow we lose access to the current version of WebPageTest and will be forced into the new version. Iβm dreading the switch. Experiments are a HUGE step backwards and I use them all the time. The interface may be consistent with Catchpoint UI, but it is boring and confusing if all you use is WPT.
22.10.2025 20:46 β π 4 π 2 π¬ 2 π 0Login flows that require me to bounce back to email are way more annoying than the traditional username/password.
22.10.2025 14:51 β π 3 π 0 π¬ 0 π 0