> (G)roff looks like it's aimed at paper outputs first

Indeed - seem to remember (?Brian) comment one journal turned down a paper due to how well typeset is was believing it'd been printed elsewhere before