Convergence: computer science

Until recently, one might have considered this immense collection of experimental information as one of the major experimentally gathered, but theoretically unexplained, bodies of data in all of the chemical sciences. As shown above, however, in recent years major progress has been made on the theory of alloys and in estabhshing empirical or semiempirical correlations for alloy phase formation this, in turn, has benefited phase diagram prediction, which remains an important goal of computational alloy theory. Two principal approaches exist, which are presently in the process of slow convergence. 

It would have been unusual to mention the importance of mathematics to physicians in a book on poisons in the eighteenth century (Richard Mead, A mechanical account of poisons in several essays, second edition published in 1708), but nearly 300 years later mathematical and computational (in silico) methods are valuable assets for toxicology as they are in many other areas of science. From such a vantage point no one would have foreseen the broad impact and importance of toxicology itself, let alone its entwined relationship with pharmaceutical and environmental research. Now is the time for an assessment of the convergence of toxicology and computational methods in these areas and to outline where they will go in the future. 

Corrosion science remains a fertile scientific endeavor, poised for advances that will benefit society. As in the past, these advances will be enabled by progress in related fields, particularly in materials characterization and computation. Indeed, an overarching observation is that the amazing recent advances in these areas portend well for the future of corrosion science as capabilities for refining time and length scales allow modeling and experimentation to converge. 

In this section, I want to address briefly the issue why the representation should be of major concern when developing a computational model. First of all, I assume (correctly or incorrectly) that the main objective of the present series of conferences is to converge on the concept of computation that would be viable not just within the classical , i.e. logical , setting but also in the natural sciences in general. 

The ESV calculations in Chapters 4 and 5 on the relatively direct D+H2 reaction demonstrate remarkably rapid convergence. This appears to be an extremely attractive computational framework for solving the four atom reactive scattering problem in full dimensionality (i.e. six dimensions with J — 0), allowing the ab initio study of many reactions of interest to science [18, 19] and technology. Indeed, such studies are now underway [20]. 

---