All I have to do is stumble upon a heated 100+ post discussion on the implementation details of an arcane feature of CSS to understand how mammoth an undertaking it must be to develop a fully compatible browser engine.
Do we need more "fully compatible" engines? I could imagine there are use cases for browser engines that work with just parts of the specification, particularly the most common ones used in the wild.
Just as a data point, the YouTube app on most TVs and devices is a stripped down browser only supporting exactly the HTML/CSS/JS needed to run the big picture client.
By tighter integration with the final product, the browser can provide specialized elements or APIs to simplify the actual application code.
That just means that specific subset of features wasn't lucky enough to make it compelling at that particular point in time for the markets it meant to address.
Change any of those variables and you may have a winning proposition.
Do we need more "fully compatible" engines? I could imagine there are use cases for browser engines that work with just parts of the specification, particularly the most common ones used in the wild.