For 30 years, early type galaxies (ETGs) have been known to have boxy and disky isophotes. While disky systems seem trivial to understand, little progress has been made in understanding boxy structures, other than the original link to pressure-supported (slow rotator) systems found by Bender in 1988. Recently the volume-limited ATLAS3D survey found no such correlation, but there are important differences between the two studies. Bender's sample contained predominantly ellipticals, while ATLAS3D's sample was dominated by lenticulars. Furthermore, Bender quoted the most extreme value of the A4 parameter in a galaxy, whereas ATLAS3D adopt the luminosity weighted value within one effective radius. Using SDSS and 2MASS imaging, we re-analyse the isophotal shape of the ATLAS3D sample using an approach motivated by Bender's method, but differing in the exclusion of boxy bulges. We recover a qualitatively similar trend to that found by Bender: boxy ETGs have lower V/Sigma while disky ETGs have higher V/Sigma. However, unlike the original result, pressure-supported systems exist with both mildly boxy or mildly disky isophote shapes and scatter equally around A4=0. Stikingly, when we consider visual morphology, the ellipticals and lenticulars behave as two different families: lenticulars rarely show boxy isophotes, while ellipticals are equally boxy or disky up to a limiting V/Sigma or Lambda; the same is not true of kinematic morphologies. We discuss the implications of these results with regard to the fast and slow rotator paradigm, and the intrinsic shape and formation histories of ellipticals and lenticulars.