In terms of switch indicator usage, am I on the right track when including notes to spatial problems with embedded Nemeth, see attachment? (I am choosing to omit the arrows)
Yes. Yes, you are on the right track when you transcribe comments in words in UEB.
I would do precisely what you’ve done in the included example, using the single-word switch indicator for a comment that is only one word but closing and then re-opening Nemeth mode for any word-based comment longer than one word.