joshwcomeau.com
Indie developer and educator. Blogging at http://joshwcomeau.com.
Previously at DigitalOcean, Khan Academy. Taught at Concordia Bootcamps.
He/Him đ
1,785 posts
21,254 followers
353 following
Regular Contributor
Active Commenter
comment in response to
post
Although I should note, itâs about a gay couple and *very* explicit in places đ
comment in response to
post
I really enjoyed Swordcrossed! www.goodreads.com/book/show/20...
comment in response to
post
Also, the State of CSS survey is currently open, so if you have a few moments, you can answer it here: survey.devographics.com/en-US/survey...
comment in response to
post
You can subscribe to my newsletter here: www.joshwcomeau.com/subscribe/
comment in response to
post
I think it depends! I agree itâs too much if this is the default behaviour for all code snippets, but if itâs a one-off flourish for a particular snippet, I think itâs great đ
comment in response to
post
Makes sense!
comment in response to
post
I didnât try running it, but I assume it would work! Itâs just a very strange thing to do đ
comment in response to
post
This is so fancy what the heck. Like a posh spaceship!
comment in response to
post
Honestly from an ROI perspective thatâs probably the right call đ I doubt the social images really make much of a difference when it comes to traffic.
But it feels like an important part, aesthetically. Like the cover of a book.
comment in response to
post
That sounds awesome!
comment in response to
post
Ah, I hadnât even heard of it!
Honestly based on Figmaâs AI web tool, Iâm pretty skeptical of any new releases they share đ
.
comment in response to
post
That would make a lot more sense than what it actually is! đ
comment in response to
post
I havenât really noticed any trends, but I also havenât really looked. Usually the topic itself has a much larger impact on how popular something will be (so my interactive guides to basic CSS stuff are my most popular, for example).
comment in response to
post
Every now and then someone asks how I make these images, expecting me to have some super-sophisticated auto-generating tool. But really thereâs no magic, I make them myself in Figma. Each one takes maybe 30 minutes, but itâs a fun lilâ task. đ
comment in response to
post
I don't really expect anyone to be able to guess correctly, but maybe yâall will give me ideas for future posts!
comment in response to
post
Typically these large numbers are caused by adding dependencies to a lockfile, but this PR doesnât add any new deps. Just a lot of demos.
comment in response to
post
A few years ago, I mightâve seen advice like this and decided to spend the day refactoring my tests, to conform with best practices.
Now I just shrug and put it into my mental library. Maybe Iâll try it on my next project, maybe not. Life will be just fine either way.
comment in response to
post
I think it depends what it is â I prefer text for most things since I can scan at my own pace (easier to scroll a blog post than scrub a video), but for some things, video makes way more sense. One example would be demonstrative things, like how to unclog a sink.
comment in response to
post
Also, I think this question is about my latest blog post, which you should check out if you havenât already!
www.joshwcomeau.com/animation/pa...
comment in response to
post
I really try to have patience for stuff like this, since I know itâs difficult to imagine what itâs like not to know something. They have all this context about what theyâre doing, but they donât provide any of it, so I have no idea what theyâre talking about. đ
But at least get my name right!
comment in response to
post
Hey Nicolas!
So I just checked, and May 2025 was ~20% lower than May 2024. I donât really think itâs because of Twitter though; hereâs what my referrers looked like in May 2024 (when I was still on Twitter):
comment in response to
post
Yeah, the contrast definitely seems pretty low! For stuff like this I think itâs fine as long as itâs configurable. Iâm guessing thereâs a "high contrast" mode in the settings.
Of course, that only helps if Apple has rigorously tested it and made sure everything still works.
comment in response to
post
Artur deserves all of the credit! I learned the trick from his work đ
And yeah, I think to do it how Apple is doing it, youâd need to use SVG filters or shaders, I doubt it can be accomplished with CSS alone.
comment in response to
post
Yeah, itâs super cool, right? đ
Just plz note that it doesnât work on iOS 16 and earlier. Thatâs like <1% of global internet traffic at this point, so not a huge group, but worth knowing all the same.
comment in response to
post
Mhmm, my headers are very boring đ love how Ahmad ties the concepts in at the top!
comment in response to
post
Itâs good to know that thatâs not clear though! Iâll see what I can do to make it clearer đ
comment in response to
post
Feedback is always appreciated! Doesnât have to start with praise â¤ď¸
In the gif you shared, you have the âpartialâ option selected, at the top. That lets you toggle between full key frames (that have the issue) and partial (that doesnât)
comment in response to
post
Hm, so when I increase the duration of the âfromZeroâ keyframe to 20s, it definitely starts bouncing in the first 25%. I think itâs more that it doesnât really have time to start bouncing at the usual speeds. Unless maybe it varies by browser?
comment in response to
post
Hm, so that seems like itâs cross fading from the two animations, but my understanding is that the twinkle opacity is effectively fading in. So the twinkle opacity is multiplied by the fade opacity.
comment in response to
post
The plan is to launch in âearly accessâ in September đ. Maybe 50% of the content will be available at launch, with the rest coming in the first half of 2026
comment in response to
post
Ack, youâre right! Set it to a constant value while debugging and forgot to revert đ
Thanks for the heads-up!