[dan-robertson]

Is there any hope of handing off keyboard shortcut customisation and recognition to the user agent instead of each web app having to reinvent them? In my ideal world it would be a requested permission (so web pages wouldn’t be able to prevent normal shortcut behaviour by default) and at some initialisation step the website would hint at the things it would like to be able to do (eg purposes, similar commands, groups, suggested keys) and the browser would figure out some reasonable mappings and offer customisability. Browsers could even offer search for web app commands (like eg Mac apps when searching from the help menu).

But web apps might not like potentially giving up so much control over their interface/UX and I think the system would need to be so complicated (eg look at Emacs with nested keymaps, remapping, multiple active keymaps, transient keymaps, and even the weird translation maps) that it would be often overkill and frequently incomprehensible and practically unconfigurable.

But the status quo where web apps often have fixed keyboard shortcuts, override browser/os defaults to do totally unrelated operations, and offer little recourse also doesn’t seem great. All I really want is for Google docs on Linux to give me Emacs (or at least macOS) style shortcuts for moving in text.