Florent Chervet has suggested on c.t.t. the following definition for a quark:
with error message "Use of \quark doesn't match its definition" if it's ever accidentally executed. I *think* this definition still fulfils the use case for quarks themselves, but I wonder if it would involve a little too much code shuffling to be worth implementing for expl3.
(And it makes it more difficult to peek inside a quark, if that's ever necessary.)
Before I forget about this idea, does anyone have any comments on the matter?