Damian
@oczki.pl
Product designer, software engineer, accessibility ally. Semantic HTML, CSS that actually cascades, and vanilla JS. Fan of puzzle games and metroidbrainias.
Almost 10 years ago, there was a discussion about this on the W3C mailing list (even with possible implementation details), but it didn't trigger any action.
lists.w3.org/Archives/Pub...
lists.w3.org/Archives/Pub...
[css-images] Negative implications of linear gradient color space choice in CSS from Mark Straver on 2016-01-22 (www-style@w3.org from January 2016)
lists.w3.org
November 6, 2025 at 11:53 AM
Almost 10 years ago, there was a discussion about this on the W3C mailing list (even with possible implementation details), but it didn't trigger any action.
lists.w3.org/Archives/Pub...
lists.w3.org/Archives/Pub...
A way to disable pre-multiplied color space in CSS, so that you have _full_ control over gradients.
Currently browsers try to read your mind when interpolating colors with alpha, and it's incompatible with some effects.
I wrote about it here: bsky.app/profile/oczk...
Currently browsers try to read your mind when interpolating colors with alpha, and it's incompatible with some effects.
I wrote about it here: bsky.app/profile/oczk...
There’s a bug in @figma.com, caused entirely by super confusing rendering of gradients in #CSS. I don't really blame Figma.
Imagine you want to show a gradient from red (at 100% opacity) to blue (at 0% opacity). In Figma, the middle is purple at 50% opacity. But CSS completely ignores the 0% color!
Imagine you want to show a gradient from red (at 100% opacity) to blue (at 0% opacity). In Figma, the middle is purple at 50% opacity. But CSS completely ignores the 0% color!
November 6, 2025 at 11:10 AM
A way to disable pre-multiplied color space in CSS, so that you have _full_ control over gradients.
Currently browsers try to read your mind when interpolating colors with alpha, and it's incompatible with some effects.
I wrote about it here: bsky.app/profile/oczk...
Currently browsers try to read your mind when interpolating colors with alpha, and it's incompatible with some effects.
I wrote about it here: bsky.app/profile/oczk...
From that image positioned directly below "Here's what to do instead", I initially thought you wanted to promote using carousels!
February 25, 2025 at 4:06 PM
From that image positioned directly below "Here's what to do instead", I initially thought you wanted to promote using carousels!
Polish (polski):
Poznaj sposoby tworzenia zachwycających interakcji z uroczymi detalami, z użyciem magii CSS, JavaScript, SVG i Canvas.
Zdradzam wszystkie moje triki!
Poznaj sposoby tworzenia zachwycających interakcji z uroczymi detalami, z użyciem magii CSS, JavaScript, SVG i Canvas.
Zdradzam wszystkie moje triki!
February 5, 2025 at 1:12 PM
Polish (polski):
Poznaj sposoby tworzenia zachwycających interakcji z uroczymi detalami, z użyciem magii CSS, JavaScript, SVG i Canvas.
Zdradzam wszystkie moje triki!
Poznaj sposoby tworzenia zachwycających interakcji z uroczymi detalami, z użyciem magii CSS, JavaScript, SVG i Canvas.
Zdradzam wszystkie moje triki!
None of the above are possible with links. The sender can modify what the link points to, fooling the initial scan by email client. Links will rot, like they always do. Linking requires me to do an action to keep the file. My email client won't scan the linked file's contents or preview them.
January 18, 2025 at 11:06 AM
None of the above are possible with links. The sender can modify what the link points to, fooling the initial scan by email client. Links will rot, like they always do. Linking requires me to do an action to keep the file. My email client won't scan the linked file's contents or preview them.
That post asks a question, but commenting is disabled - is that intended?
I vastly prefer attachments, because files are stored on the end I can control. No risk of link rot. The sender can't modify the file after sending (good for legal documents). Email client scans them and allows previews.
I vastly prefer attachments, because files are stored on the end I can control. No risk of link rot. The sender can't modify the file after sending (good for legal documents). Email client scans them and allows previews.
January 18, 2025 at 11:06 AM
That post asks a question, but commenting is disabled - is that intended?
I vastly prefer attachments, because files are stored on the end I can control. No risk of link rot. The sender can't modify the file after sending (good for legal documents). Email client scans them and allows previews.
I vastly prefer attachments, because files are stored on the end I can control. No risk of link rot. The sender can't modify the file after sending (good for legal documents). Email client scans them and allows previews.
Does this work (even via cronjob) when you're in Deck's game mode?
There are games that like to crash and wipe your save (e.g. ZTD), so I'd love a hourly backup of specific titles, even during gameplay, without leaving game mode.
There are games that like to crash and wipe your save (e.g. ZTD), so I'd love a hourly backup of specific titles, even during gameplay, without leaving game mode.
January 16, 2025 at 1:00 PM
Does this work (even via cronjob) when you're in Deck's game mode?
There are games that like to crash and wipe your save (e.g. ZTD), so I'd love a hourly backup of specific titles, even during gameplay, without leaving game mode.
There are games that like to crash and wipe your save (e.g. ZTD), so I'd love a hourly backup of specific titles, even during gameplay, without leaving game mode.
Yeah, every case is different. I like to mention that "delete repository" pattern when teaching about accessibility and preventing huge mistakes.
Thanks for the discussion!
Thanks for the discussion!
January 13, 2025 at 11:43 AM
Yeah, every case is different. I like to mention that "delete repository" pattern when teaching about accessibility and preventing huge mistakes.
Thanks for the discussion!
Thanks for the discussion!
Reserving space is tricky when you consider users who changed their root font size, need to support translations (e.g. German can be much longer than English), or have an error message that just can't be conveyed in a few words and needs more lines.
January 13, 2025 at 11:30 AM
Reserving space is tricky when you consider users who changed their root font size, need to support translations (e.g. German can be much longer than English), or have an error message that just can't be conveyed in a few words and needs more lines.
What I've learned from @adamsilverhq.bsky.social's Form Design Mastery course is that if you validate on submit, layout shifts aren't an issue. Users know how to scroll. And it would be best to break up a form into smaller parts if it has that many fields!
January 13, 2025 at 10:50 AM
What I've learned from @adamsilverhq.bsky.social's Form Design Mastery course is that if you validate on submit, layout shifts aren't an issue. Users know how to scroll. And it would be best to break up a form into smaller parts if it has that many fields!
Even better - display hint text right below the label, above the input field, so that the text isn't covered by browser's autocomplete popup or a phone's virtual keyboard. Same for error messages.
January 13, 2025 at 9:30 AM
Even better - display hint text right below the label, above the input field, so that the text isn't covered by browser's autocomplete popup or a phone's virtual keyboard. Same for error messages.
Here's some research on it: www.researchgate.net/publication/...
(PDF) What's the Most Effective Way to Present Two-State Toggle Buttons?
PDF | Two-state toggle buttons (e.g., On/Off, Yes/No, etc) are very common on web sites and in mobile apps. Both the Android and iOS toolkits have... | Find, read and cite all the research you need on...
www.researchgate.net
January 10, 2025 at 4:19 PM
Here's some research on it: www.researchgate.net/publication/...
Personally, if I really, really (really) have to use a segmented control, I insist on only using it if there are 3+ options to choose from, and I'm 100% sure people will know which option is selected.
I've seen way too many implementations with 2 options where it's not clear which one is selected.
I've seen way too many implementations with 2 options where it's not clear which one is selected.
January 10, 2025 at 2:07 PM
Personally, if I really, really (really) have to use a segmented control, I insist on only using it if there are 3+ options to choose from, and I'm 100% sure people will know which option is selected.
I've seen way too many implementations with 2 options where it's not clear which one is selected.
I've seen way too many implementations with 2 options where it's not clear which one is selected.
Sad news, but instability is expected at this early stage of WCAG3.
APCA was quite a big improvement, and there are lots of a11y-related websites that assumed APCA is the future. Are there any plans to make a semi-official statement, so that this info gets around?
APCA was quite a big improvement, and there are lots of a11y-related websites that assumed APCA is the future. Are there any plans to make a semi-official statement, so that this info gets around?
January 10, 2025 at 7:36 AM
Sad news, but instability is expected at this early stage of WCAG3.
APCA was quite a big improvement, and there are lots of a11y-related websites that assumed APCA is the future. Are there any plans to make a semi-official statement, so that this info gets around?
APCA was quite a big improvement, and there are lots of a11y-related websites that assumed APCA is the future. Are there any plans to make a semi-official statement, so that this info gets around?