Material-UI is releasing v5 and rebrands to mui.com

Hi HN. I'm one of the creators of MUI. The whole team is very excited to be sharing the results of the last two years of our work!

The last time we were on HN was 7 years ago with https://news.ycombinator.com/item?id=8582439! Back then, the community got hooked up in the integration of two shiny new trends: React and Material Design. A lot has changed in the React community since then.

In v5, we have focused on improving the customization experience, making the components as flexible as possible. You can find the following improvements:

- We have a new rebrand that distinguishes ourselves from Material Design: https://mui.com/blog/material-ui-is-now-mui/.

- We have moved from JSS to emotion, using the styled-components API as a primitive. This was a highly requested change from the community: https://github.com/mui-org/material-ui/issues/6115

- We have introduced styling props, and a sx prop: https://mui.com/system/basics/

- We have a growing set of alpha unstyled components, e.g. https://mui.com/components/autocomplete/#useautocomplete

- We have incorporated a company, sustaining over 10 people.

- We have introduced an advanced data grid: https://mui.com/components/data-grid/demo/

Thanks!

When I read this quicly I thought 'Material-UI is merging with Xiaomi's MIUI [1]' - and I suspect I will not be the only one making this mistake...

[1] https://home.miui.com/

Reading through the styling part and working with mui4 and other widget libraries in the past - and seeing the 5 solutions listed that basically do the same thing - holly shit how did it get to this point...

I got the value of JSX even if it felt messy - but the css-in-js becoming popular on frontend is one of the why I switched to full backend roles now.

And when he says that changing the implementation of a styling solution improves component performance 5x ... at what point is it OK for a fancy styling abstraction layer that serves no real purpose (it was just as shit to customise as any other solution, doubt it made people significantly more productive, and they said it was a maintenance burden to maintain abstraction layers) suck up that much performance ? Even the "fast" path is order of magnitude slower compared to good old CSS...

There's a codemod [1] to automate most of the code changes needed, and a comprehensive migration guide [2].

[1] https://mui.com/guides/migration-v4/#preset-safe

[2] https://mui.com/guides/migration-v4/

Feedback so far has been positive...