iiinova — where your digital self keeps evolving

CSS Box Sizing Module Level 4

Does that work? The spec suggests the ICB of the iframe's document becomes locked. Is that right @bram.us? drafts.csswg.org/css-sizing-4...

drafts.csswg.org

230

So, with this new API, what do I do if the user resizes the main window, such that the iframe element changes width?

And the parent page will never be resized?

000

Isn't it, like, less than 10 lines of code?

210

How do I call that from the parent document? I'm talking about cases where the iframe has changed width on the parent.

020

Right, and if the iframe is width:100%, and therefore changes width due to things in the parent page, how does the parent page tell it to update height? (cc @bram.us)

140

Having to use JS to update the iframe's size in response to simple things like opening a <details> inside the iframe seems a bit useless, no?

Cyd Stumpel @cydstumpel.nl · 16/08/2026

Reposted by Jake Archibald

Started working on my demo for Fronteers Dark Mode, this is built with CSS scroll-driven animations, no JS 💅 My talk will go into detail on where you can replace JS with CSS! Tickets are still available: fronteersconf.org This cool effect is inspired by @thomasaufresne.com → bennettandclive.com

7898

Ohh, so there's value in a gain map even for 'HDR' screens, because it can adapt better to the screen & settings (eg brightness)?

210

Yeah, I'm not sure what's going on with the PQ AVIF in Firefox. That seems like a bug to me. But even the PQ JXL… it's not awful, but with the gainmap I get full control over how the SDR version appears.

000

random-stuff.jakearchibald.com

It depends how the HDR values get crushed down to SDR. With gain maps, you get to choose both the SDR and HDR representations (just as you would with two separate images). Compare PQ AVIF to SDR-based gain map on an SDR screen:

HDR formats

100

It's just a JPEG with a gain map. It's well supported I think. But with JPEG you're still starting with an 8-bit image.

020

cjxl doesn't support gain maps yet, and neither does jxl-rs, but it'll be interesting to see if it has this same overhead problem. I suspect it will.

130

MDN Web Docs @developer.mozilla.org · 25/09/2026

I've been playing around with HDR images on the web. It seems that, for AVIF (the only web format that supports high bit depth and gain maps), gain maps add a lot of overhead. On the web it'd be much better to use <picture> to serve the right image for the screen.

2141

Reposted by Jake Archibald

🆕 Customize MDN's browser compat tables! Press ⚙️ Settings on any compat table to choose which browsers and engines you see, and hide the ones you don't need. Your picks are saved in your browser. Try it 👇 developer.mozilla.org/en-US/docs/W...

MDN's browser compatibility table for the HTML popover attribute. Its columns cover desktop and mobile browsers, showing which versions support the attribute and its "hint" value. A "Settings" button sits above the table.

The Browser compatibility settings dialog on MDN. Browsers appear as chips grouped under Desktop, Mobile, Server and XR, and each one can be added to or removed from the table.

27725

Reposted by Jake Archibald

Niels Leenheer @html5test.com · 24/09/2026

Fronteers Dark Mode conference is next week! Monday I'll send badges to the printer... which means if you want a proper badge... sign up now! We've got a handful of tickets left and they are only €100 for an evening of front-end talks by some amazing speakers. www.fronteersconf.org

fronteersconf.org

Fronteers Dark Mode, Friday, October 2nd, Dordrecht

Join us at Cinema de Witt in Dordrecht for our after work conference.

098

Jake Archibald @jakearchibald.com · 24/09/2026

I don't understand. Can you… clarify or reword this?

101

A picture of a puppy is nice. I feel you're avoiding talking about something that's representative of the final image, because blurhash is not.

100

Creating something unrecognisable is really really easy. But what problem does 'unrecognisable' solve?

200

Given that we already have sold colours and gradients, and there's clearly the ability to do a more structured preview in ~500 bytes, what's the justification for something like a blurhash? What use-case is it solving?

Both X and WhatsApp use more representative image previews. The smallest thing that fills the empty area instantly is a solid colour and this is already possible. The platform already has gradients too. I don't see the justification for a new inefficient gradient format.

Keith Cirkel @keithamus.social · 22/09/2026

Reposted by Jake Archibald

I just released avash: www.keithcirkel.co.uk/avash/ It’s like blurhash but uses av1 frame data encode an image, resulting in a (typically) more readable image in the same bytes as a blurhash. github.com/keithamus/av...

keithcirkel.co.uk

avash encoder

3202

Jake Archibald @jakearchibald.com · 22/09/2026

Something being popular isn't reason enough, especially given that apps with more resources seem to be moving away from blurhash. Take your example. I guessed the image might be a seascape. It wasn't. Compare your example to a 1 kB AVIF:

A blur that's vaguely blue up top, then kinda red in the lower middle, and black in the bottom left.

A blur that clearly shows a blue sky with some clouds, what looks like a building in the lower left, and a pyramid-shaped thing in the lower middle.

Jake Archibald @jakearchibald.com · 22/09/2026

I don't see what use-cases they serve vs a regular gradient (and something we could expand to a mesh gradient later). Are the use-cases you're thinking of documented anywhere?

100

It's less about scope and more about where things end up getting fully computed. Without @property, or with a * syntax, it doesn't convert things like sibling-index() to their number, so that part is left until you use it with var(). With @property it fully computes in the custom property.

Controlling when CSS custom property values are computed

This technique can be used to read em & container query units for a particular element, then use them in a child element - I just used it to allow an element to generate negative margins to 'break out' of its container. Claude told me it wasn't possible until I gave it the link 😁

jakearchibald.com

The default might not be what you expect, and it has a big impact on your styles.

3352

Really excited about this one! A new web conference in Liverpool, UK.

0401

Yeah. I haven't dug enough to put anything in the "pro" column, but having to dig for it is an issue. Whereas there's a big obvious thing in the "con" column.

Jake Archibald @jakearchibald.com · 20/09/2026

There's lots I haven't yet understood about atproto, but I simply won't do the standard.site thing whilst it offers a worse UX than… not doing it.

5635

Jake Archibald @jakearchibald.com · 20/09/2026

I am delighted that person is upset.

2120

Jake Archibald @jakearchibald.com · 19/09/2026

I think we have some failing tests with that combination

000

Jake Archibald @jakearchibald.com · 18/09/2026

Now I've come back to it, I guess that demo makes total sense, but it cost me time trying to figure out what was going on with that first line.

100

Jake Archibald @jakearchibald.com · 18/09/2026

Yeah, I'm working from Figma. I've ran into some surprising behaviours with text-box, I really thought this one was a bug, but it seems to be cross-browser random-stuff.jakearchibald.com/interesting-..., so I'm not sure it should be in a reset. But I've used it a lot.

random-stuff.jakearchibald.com

Interesting text-box side-effect

141

Jake Archibald @jakearchibald.com · 18/09/2026

I'm building a site where I can use pretty much any baseline CSS feature and… ahhh @scope and text-box are so good.

4591

Reposted by Jake Archibald

Niels Leenheer @html5test.com · 18/09/2026

Two weeks from now Fronteers Dark Mode will happen in Dordrecht in The Netherlands. There are still a few tickets left. Be quick ⚠️ Join us for an afternoon and evening of great talks. With a spectacular line up: @anjana.dev, @tunetheweb.com, @cydstumpel.nl, @csswizardry.com and @heydonworks.com

21310

Jake Archibald @jakearchibald.com · 17/09/2026

I remember a conference after party a decade+ ago where @johnresig.com was stood in the corner on his own because everyone (including me) was too shy to speak to him. Fwiw, if you ever want to talk to me at one of these things, please just do it. I'm also shy.

2713

Reposted by Jake Archibald

Viewport @viewportconf.org · 03/09/2026

We just finished the beta of our website. Look out for updates on Viewport soon! You can sign up to our mailing list to be kept in the loop. viewportconf.org

viewportconf.org

Viewport 2027 — An independent conference about the web

An independent conference about the web, the people who build it, and the people we build it for. Liverpool, May 2027.

1136

Jake Archibald @jakearchibald.com · 15/09/2026

For high density, is it just a matter of targeting a lower SSIMU2, or do we need a new algorithm for high density?