fixes typo for in-page link (#1720)

HTTPArchive · Dec 9, 2020 · 1b1ca85 · 1b1ca85
1 parent 130aece
commit 1b1ca85
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/src/content/en/2020/css.md b/src/content/en/2020/css.md
@@ -202,7 +202,7 @@ There is always a lot of inertia when we change something in the Web platform af
 
 By far the most popular pseudo-classes are user action ones, with `:hover`, `:focus`, and `:active` at the top of the list, all used in over two thirds of pages, indicating that developers like the convenience of specifying declarative UI interactions.
 
-`:root` seems far more popular than is justified by its function, used in one third of pages. In HTML content, it just selects the `<html>` element, so why didn't developers just use `html`? A possible answer may lie in a common practice related to defining custom properties, [which are also highly used](#custom properties), on the `:root` pseudo-class. Another answer may lie in specificity: `:root`, being a pseudo-class, has a higher specificity than `html`: (0, 1, 0) vs (0, 0, 1). It is a common hack to increase specificity of a selector by prepending it with `:root`, e.g. `:root .foo` has a specificity of (0, 2, 0) compared to just (0, 1, 0) for `.foo`. This is often all that is needed to nudge a selector slightly over another one in the cascade race and avoid the sledgehammer that is `!important`. To test this hypothesis, we also measured exactly that: how many pages use `:root` at the start of a descendant selector? The results verified our hypothesis: a remarkable 29% of pages use `:root` that way! Furthermore, 14% of desktop pages and 19% of mobile pages use `html` at the start of a descendant selector, possibly to give the selector an even smaller specificity boost. The popularity of these specificity hacks strongly indicates that developers need more fine grained control to tweak specificity than what is afforded to them via `!important`. Thankfully, this is coming soon with [`:where()`](https://developer.mozilla.org/en-US/docs/Web/CSS/:where), which is already [implemented across the board](https://caniuse.com/mdn-css_selectors_where) (albeit behind a flag in Chrome for now).
+`:root` seems far more popular than is justified by its function, used in one third of pages. In HTML content, it just selects the `<html>` element, so why didn't developers just use `html`? A possible answer may lie in a common practice related to defining custom properties, [which are also highly used](#custom-properties), on the `:root` pseudo-class. Another answer may lie in specificity: `:root`, being a pseudo-class, has a higher specificity than `html`: (0, 1, 0) vs (0, 0, 1). It is a common hack to increase specificity of a selector by prepending it with `:root`, e.g. `:root .foo` has a specificity of (0, 2, 0) compared to just (0, 1, 0) for `.foo`. This is often all that is needed to nudge a selector slightly over another one in the cascade race and avoid the sledgehammer that is `!important`. To test this hypothesis, we also measured exactly that: how many pages use `:root` at the start of a descendant selector? The results verified our hypothesis: a remarkable 29% of pages use `:root` that way! Furthermore, 14% of desktop pages and 19% of mobile pages use `html` at the start of a descendant selector, possibly to give the selector an even smaller specificity boost. The popularity of these specificity hacks strongly indicates that developers need more fine grained control to tweak specificity than what is afforded to them via `!important`. Thankfully, this is coming soon with [`:where()`](https://developer.mozilla.org/en-US/docs/Web/CSS/:where), which is already [implemented across the board](https://caniuse.com/mdn-css_selectors_where) (albeit behind a flag in Chrome for now).
 
 {{ figure_markup(
   image="popular-selector-pseudo-classes.png",