Hi,

I think we could possibly do that as a tool tip, although it is somewhat mitigated in recent NVDA alpha snapshots where a form of context-sensitive help was added to NVDA interface controls (at the moment if you press F1 from NVDA settings controls, the relevant section from the user guide will open). The next step would be generalizing this initial version of context-sensitive help to other places, but that will involve reimagining the current approach as well.

For a practical demonstration of input composition settings, take a look at Welcome to NVDA tutorial set, which can be found at:

https://www.josephsl.net/tutorials

Cheers,

Joseph