The ability of some liquids to vitrify during supercooling is usually seen as a consequence of the rates of crystal nucleation (and∕or crystal growth) becoming small [D. R. Uhlmann, J. Non-Cryst. Solids 7, 337 (1972)]--and thus a matter of kinetics. However, there is evidence dating back to the empirics of coal briquetting for maximum trucking efficiency [D. Frenkel, Physics 3, 37 (2010)] that some object shapes find little advantage in self-assembly to ordered structures--meaning random packings prevail. Noting that key studies of non-spherical object packing have never been followed from hard ellipsoids [A. Donev, F. H. Stillinger, P. M. Chaikin, and S. Torquato, Phys. Rev. Lett. 92, 255506 (2004); A. Donev, I. Cisse, D. Sachs, E. A. Variano, F. H. Stillinger, R. Connelly, S. Torquato, and P. M. Chaikin, Science 303, 990 (2004)] or spherocylinders [S. R. Williams and A. P. Philipse, Phys. Rev. E 67, 051301 (2003)] (diatomics excepted [S.-H. Chong, A. J. Moreno, F. Sciortino, and W. Kob, Phys. Rev. Lett. 94, 215701 (2005)] into the world of molecules with attractive forces, we have made a molecular dynamics study of crystal melting and glass formation on the Gay-Berne (G-B) model of ellipsoidal objects [J. G. Gay and B. J. Berne, J. Chem. Phys. 74, 3316 (1981)] across the aspect ratio range of the hard ellipsoid studies. Here, we report that in the aspect ratio range of maximum ellipsoid packing efficiency, various G-B crystalline states that cannot be obtained directly from the liquid, disorder spontaneously near 0 K and transform to liquids without any detectable enthalpy of fusion. Without claiming to have proved the existence of single component examples, we use the present observations, together with our knowledge of non-ideal mixing effects, to discuss the probable existence of "ideal glassformers"--single or multicomponent liquids that vitrify before ever becoming metastable with respect to crystals. We find evidence that "ideal glassformer" systems might also be highly fragile systems, approaching the "ideal glass" condition. We link this to the high "volume fragility" behavior observed in recent hard dumbbell studies at similar length∕diameter ratios [R. Zhang and K. S. Schweitzer, J. Chem. Phys. 133, 104902 (2010)]. The discussion suggests some unusual systems for laboratory study. Using differential scanning calorimetry detection of fusion points T(m), liquidus temperatures T(l), and glass transition temperatures T(g), we describe a system that would seem incapable of crystallizing before glass transition, i.e., an "ideal glassformer." The existence of crystal-free routes to the glassy state will eliminate precrystalline fluctuations as a source of the dynamic heterogeneities that are generally considered important in the discussion of the "glassy state problem [P. W. Anderson, Science 267, 1615 (1995)]."