[dmitriid]

`className` to `class` is a bad decision.

React is a very thin layer on top of JS. `className` is JS. Specifically, it comes directly from DOM APIs. Changing it to `class` (while retaining all other DOM API-compatible prop names) is a really terrible decision that relies on “the wisdom of the crowds”.

[MBCook]

It’s JS, but not HTML.

And most of the time you see it you’re making JSX elements, where you would expect to use the word ‘class’.

I think they’re right it’s more consistent with what people would expect to happen.

The fact there is a special note in the docs calling out that everyone gets it wrong is a sign it was a problematic choice.

[quink]

JSX attributes aren't HTML attributes. One way you can tell is because they don't have any namespace whatsoever.

Similarly when you add a style to an element you don't get to use `background-color` for instance, it's `backgroundColor`. Yet they're not revisiting that decision, even though it's exactly the same, i.e. `background-color` can't be used to to a JS syntax conflict.

It seems arbitrary to make JSX attributes pertaining to a DOM that don't have a namespace match XML attributes in a markup language... while not at the same time doing the same to the CSS object model property names for instance.

[pests]

The backgroundColor vs background-color thing has nothing to do with React.

That is more JSS which is one attempt at CSS-in-JSS. The removed hyphen is needed because JSS requires property names as object property names where hyphens can be problematic.

Other CSS-in-JSS libraries, such as emotion.sh, has support for CSS (vs JSS above) inside of JS, including React components, and can keep default CSS syntax. Example from the emotion.sh homepage:

  const Link = styled.a`
    min-width: 12rem;
    margin: 0 auto 20px;
    padding: ${props => props.primary ? 18 : 16}px;
    border-radius: 5px;
    text-decoration: none;
    border: ${props =>
      props.primary ? 'none' : '3px solid currentColor'};
    background: ${props =>
      props.primary &&
      'linear-gradient(90deg, #D26AC2, #46C9E5)'};
    color: ${props =>
      props.primary ? '#1D2029' : '#D26AC2'};
    &:hover {
      opacity: 0.95;
    }
    @media (min-width: 768px) {
      margin: 0 20px 0 0;
      &:last-child {
        margin: 0;
      }
    }
  `

If you notice variables and props are available via string interpolation. emotion.sh also makes inline styling easier:

  render(
    <div
      className={css`
        background-color: hotpink;
        &:hover {
          color: ${color};
        }
      `}>
      This has a hotpink background.
    </div>
  )

[edited for formatting and word choice]

[quink]

JSS? `<div style={{backgroundColor: 'black'}} />` is standard React / ReactDOM.

> The backgroundColor vs background-color thing has nothing to do with React.

You’re right, but it absolutely has to do with ReactDOM which is being discussed here.

Which, btw, is called ReactDOM not ReactHTMLandCSSandStuff.

Although reading Dan Abramov’s further comment ReactDOM is a bit of a misnomer, but from an API perspective it’s definitely more DOM especially with the CSS precedent and lack of XML namespaces vs. the `aria` attributes being the only major evidence to the contrary. I mean, sure it only sets `x` properties directly, but as an end user of the library what do I care about the internal implementation?

Edit: A new `classList` property that accepts an array would be nice as that way anyone using `className` has a superior API to migrate to and stay away from this change entirely. It’s not like it’s setting it directly so just convert to a DOMTokenList on the fly.

[pests]

Oops I totally forgot about inline styling like that without a preprossor. You are correct.

Although it does seem like ReactDOM already does some weird stuff already in regards to vendor prefixing.

[dmitriid]

Exactly. It’s JS and not HTML.

Because in JS that attribute/property is `className`. Because `class` is a reserved name in JS.

Same goes for `htmlFor`, for example. So now you will have a situation where all props directly correspond to DOM APIs, and one prop that isn’t. A strange choice given how the stated goal is to be more consistent with DOM APIs.

[angersock]

And yet the entire reason for using JSX in the first place is that it looks like HTML. This is a reasonable change.

[danabramov]

So I wrote in detail why I think `class` is the right choice here. Hope it helps.

https://github.com/facebook/react/issues/13525#issuecomment-...

[dmitriid]

Leaving my comment here as well for increased visibility:

Current state of things:

- React is inconsistent

- API is camelCase

Proposed state of things:

- React is inconsistent

- API is camelCase

- `className` -> `class`

Perceived benefits:

> If React was open sourced today, it seems like the pros of allowing class (closer conceptually to what most people expect, less typing for the most commonly used prop) outweigh the downsides (slightly more typing to intercept it — in which cases you'll probably just want the spread operator anyway).

Actual downsides:

- the entire ecosystem that depends on `className` stops working (huge upgrade effort)

- the entire vast body of books, tutorials, examples, code, posts, articles becomes ever so slightly invalid

- the only valid JS remaining is additional typing: `const {class: cls} = props`. Every other possible use-case in plain JS becomes invalid

- React API remains inconsistent, and breaks further assumptions (why `htmlFor` not `for` etc.)

If I were a product manager, my immediate reaction to the change would be: wat?

[danabramov]

It will be fully consistent if we change `for` too (and a couple of rare SVG attributes).

[chrisweekly]

Highly recommend reading Dan's reasoning; it makes a lot of sense. This quote is as close as I can get to distilling it:

>"I think this is the primary reason it's desirable for props like tabIndex, cellSpacing, and most other DOM-related props to follow the camelCase convention. It's not because they're DOM property names. It's because they often end up in component APIs. And we want component APIs to be consistently camelCase."

[wLMeOv2qjsFOjZG]

But both class and className are camelCased.

[grandpoobah]

Your argument is addressed by Dan Abramov in the comments on the Github issue -

https://github.com/facebook/react/issues/13525#issuecomment-...

tl;dr, There's already a lot of inconsistency, this is not a good argument.

That said, I don't like the className->class change either. It's a reserved keyword, it always made perfect sense why they originally went with className.