While updating the TodoMVC applications, we noticed a few a11y related issues.
For example inconsistency in using inputs with and without labels, or using inputs without a form and therefore having to manage keyboard events more extensively.
These inconsistencies don't interfere with the test results, nor do they speed up the workloads once addressed most likely.
That said, I believe that since our workloads are supposed to represent apps on the web, we should also advocate for best practices where it makes sense. This will hopefully also encourage other engineers to follow similar patterns when developing their own applications.
While updating the TodoMVC applications, we noticed a few a11y related issues. For example inconsistency in using inputs with and without labels, or using inputs without a form and therefore having to manage keyboard events more extensively.
These inconsistencies don't interfere with the test results, nor do they speed up the workloads once addressed most likely.
That said, I believe that since our workloads are supposed to represent apps on the web, we should also advocate for best practices where it makes sense. This will hopefully also encourage other engineers to follow similar patterns when developing their own applications.