Nice analysis Steve! I noted under the "Screen Readers and text formatting" section, you had a note that "The most configurable is JAWS via speech and sound schemes". Is there anything particular you'd like to see added in NVDA in terms of configurability here? It would definitely be worth an issue on our tracker at https://github.com/nvaccess/nvda/issues if you can put forward a specific recommendation (or if you give me enough detail, I can write it up)