Hi Richard, I had few questions pop in my mind reading the doc: * What does the n prefix mean? What's the mnemonic? * These new `->` and `->>` operators are JSON only? That's why you moved JSON into the Core? * Do the JSON functions use a subtype? Such that nested JSON functions can know whether the argument is already validated well-formed JSON? About the new N functions, I don't think they are for me, 'cause I tend to be on the strict side of schemas, and would have a CHECK constraint with `json_valid()`, if I expect a column to be JSON. But in the long tradition of SQLite being *flexibly typed*, I can see where they'd simplify the SQL sometimes. Regarding the *syntax sugar* operators `->` and `->>`, why not. I'm surprised, they do not seem to *fit* the SQLite moto of lite/light, and but if JSON becomes Core, why not again. I've seen these operators in the PG doc, and they do not aid readability of the SQL IMHO, especially since there are tons such operators in PG.