Re: [w3ctag/design-reviews] Other Spec Review: [css-text] `text-fit` property (Issue #1208)

tkent-google left a comment (w3ctag/design-reviews#1208)

Thank you for the review and feedback!

### 1. DX and testing layouts
While the rendered font size adapts to the available inline space, the layout calculation itself is deterministic for a given container size, text content, and font. 

Currently, authors achieve this effect using JavaScript workarounds (e.g., `ResizeObserver` combined with DOM measurement loops), which are prone to timing issues and FOUC. Providing a declarative CSS property improves DX and makes it easier for authors to inspect and test layouts across different viewport/container sizes using standard browser DevTools and automated layout tests.

### 2. Latest status on CSSWG `#12886` and text scaling
Yes, [w3c/csswg-drafts#12886](https://github.com/w3c/csswg-drafts/issues/12886) represents the primary tracking thread on the CSSWG side. In addition to that thread, we also discussed a potential zoom heuristic with the ARIA WG in February ([2026-02-19 ARIA WG minutes](https://www.w3.org/2026/02/19-aria-minutes.html)).

In those discussions, we explored a heuristic based on `max(<post-fit size at 100% zoom> * <zoom factor>, <computed font-size> * <zoom factor>)`. However, it has a performance cost (requiring multiple layouts) and there was no clear consensus yet on the best path forward across the groups. Your point about text becoming *too large* when combining page zoom with full-screen magnification is a great point—a maximum scaling limit (as discussed in `#12886`) could potentially help address the concern.

### 3. UA toggle and fingerprinting
We agree that a user-facing UA toggle would expose a fingerprinting surface, as pages could detect whether `text-fit` is active by measuring resulting box or line dimensions via JavaScript (similar to existing browser "minimum font size" settings). Addressing the legibility concerns at the spec level (such as via scaling limits in `#12886`) may be preferable if it avoids introducing a new fingerprinting vector. We look forward to hearing any further insights from APA's ongoing discussions.


-- 
Reply to this email directly or view it on GitHub:
https://github.com/w3ctag/design-reviews/issues/1208#issuecomment-6570371170
You are receiving this because you are subscribed to this thread.

Message ID: <w3ctag/design-reviews/issues/1208/6570371170@github.com>