Understanding the structures and physical properties of materials, such as hardness and the occurrence of metalinsulator transitions, has long been a goal of the chemist and materials scientist alike. Our understanding of the structures of molecules and solids has progressed dramatically in the past decade driven by computational advances in terms of both hardware and basic theory. In the area of the solid state especially, this period has seen the coming together of “chemical” and “physical” approaches to solids.1-3 This period has witnessed some remarkable successes. The ability to study a wide range of complex materials using the tight-binding with overlap, or extended Hu¨ckel, approach has been well-established.2 The success of tight-binding calculations to study complex systems by employing the technique of second moment conservation is an important advance.4 Importantly, over this period there has been a transfer of tight-binding technology from the physical theorist to chemical theorist and in turn to the practicing solid-state chemist. In terms of numerically accurate methods, the ability from pseudopotential-based calculations to predict the transition pressure from one structure to another5 and the prediction of properties such as hardness are milestones. The local density approximation6 has provided for solids an effective way to estimate many-body effects. By using generalized gradient corrections (GGC), it has been possible with full potential LAPW calculations7 to effectively probe the factors influencing the spin state of ions in solids or, more generally, the magnetic or localization behavior of electrons. (See ref 8 for a succinct description of the present state of this rapidly advancing field.) The LMTO method9 is proving to be an extremely useful tool with which to study the electronic structure of solids, especially close-packed ones. (It is regarded by many hard-line theorists as an “old-fashioned” approach but has many advantages for the structural chemist.) Undoubtedly, the dramatic change in computational power over the past few years has accelerated the application of such methods to “real” systems. This has brought its problems, however. We now have even more numbers. The question of how to organize them and develop new insights and models of use to the chemist has certainly become a challenge.