Thread by @gregwhitworth

I appreciate folks asking the value of a potential new HTML element for checkbox as it seems so simple to build and make them accessible. But in doing research we found that out of 101 different checkboxes across the Interbrands top 100 brands and found 44% have an issue.

* Invisible focus state: 16%
* Non-interactive label: 30%
* No computed name: 59%

Now a non-interactive label is not inherently wrong but it does produce and incoherent UX for users. But the other 2 are pretty much non-starters for a successful user experience in any scenario.

You can read a bit more here: open-ui.org/components/che…

Note: We didn't list the sites specifically here on purpose to not do a shame game but go ahead and take an hour and go browse the web and look at checkboxes to see if they have any of the above issues.

One thing that we didn't document but should have added in was the complete lack of any label. Most of these sites didn't have this issue but I'm sure the long tail would.

Github issue regarding a potential new checkbox element is here: github.com/openui/open-ui…

Your opinions are welcome

Oh and one other thing as I denote in the issue, this doesn't get into the scenarios that want complex content which currently you'd need to reach for a custom implementation with a hidden <input type="checkbox"> to accomplish.

At any rate, maybe I'm wrong and we don't need it. One thing that I do think we'll need no matter what is a grouping component of some form for radio and checkbox based on investigations in this space.