I might actually put the limits after the integral sign. Then it makes an indefinite integral look like a partially-applied function just waiting for a tuple of limits to make it a definite integral. That might get hairy for iterated integrals, besides being highly unconventional.
Yeah, that would seem to be more practical with respect to currying (and perhaps dynamic dispatch). I was just trying to make it look as similar as possible to the Leibniz notation. :)
