Nice attempt at a "proof by intimidation" but I would see it as an invitation to pick a non-integer! 🙂

If you change the numbers to 1 and 30 (and explicitly ask for an integer) then I doubt that 1 and 30 would be as likely outcomes as the other integers in the range.

But my main point was to point out the ambiguity, not that one choice was necessarily more correct than the other.

I totally agree with you that there are worse "features" in SQL (and many other languages) and with Hugo that complaining is a consequence of caring...

(edit: I would not even calls this a 'bug', like I would call certain other "features", but simply a 'bad choice')