Follow

Reading over Seirdy's "Best practices for inclusive textual websites", some webstandards he mentions that I am supporting or plan to support in my browserengines relating to performance, accessibility, & privacy:

* data: URIs
* TLS 1.3
* caching headers
* I can be talked into CSS Containment, a prerequisite for Container Queries
* Darkmode (and alternate styles)
* Both web & native fonts
* I don't do preloading, lazy-loading, or CSP.
* Basically all HTML5 elements

1/2

Some nuances when it comes to performance in my work:

I don't implement progressive HTML loading, which makes some performance tips moot. Not worth the effort...

If you use media queries to respond to window resizing, that does increase the amount of effort my software needs to go through to relayout the page.

If I treat webpages as documents under Orca, it appears I won't need to add performance overhead.

2/2

The link again: seirdy.one/posts/2020/11/23/we

@alcinnz I don't think Containment is necessary for Container Queries; CSS containers and CSS containment are different concepts with unfortunate naming. Containment is useful for pages with thousands of DOM nodes, like long Wikipedia articles. If your browser can handle them easily as-is then it's no biggie.

It's also useful for apps to indicate that state changes are confined, to improve performance by limiting DOM calculations. If you plan to ignore JS then that's not a relevant concern.

What containment does NOT address is perf of re-calculating the accessibility tree when the DOM changes: https://seirdy.one/posts/2020/11/23/website-best-practices/#performance-of-assistive-technologies
Again, barely an issue if you ignore JS.

@Seirdy Hearing from the spec writers, it is a prerequisite. The issue with Container Queries is that without some strategy styling an element based on its size may result in an infinite loop. Which I'm opposed to even more than other browserdevs!

Containment provides a way to break this loop, and is why I'd be happy to implement Container Queries if there's enough desire for it.

@Seirdy O.K., I can't say whether I'm more opposed to looping in CSS. But I am strongly opposed, I'm trying not to make Turing Complete!

Sign in to participate in the conversation
FLOSS.social

For people who care about, support, or build Free, Libre, and Open Source Software (FLOSS).