The Community for Technology Leaders

Computational Science: We'll Know It When We See It


Pages: p. 2

A mathematician friend recently began taking the whole sequence of undergraduate courses in chemistry and physics partly as a hobby and partly to learn about how the universe actually works. Some of the courses appear to be current versions of the ones I took as an undergraduate a few decades ago. (Digression: it's time to start computing dates in "Jack Benny time units." In such units, my courses were a few months ago.) Several things about these courses struck me as worthy of note. At first glance, some of the topics in the chemistry courses looked to me like the ones I'd had in physics courses. But I checked my old textbooks and found that I was wrong. The basic notions of quantum theory and statistical mechanics have been part of the undergraduate chemistry curriculum for a long time.

Once I'd started rummaging through old texts, it was hard to stop. That's why it's important not to look at anything when trying to clear out old files and papers. If you look at it, it becomes a treasure. Anyhow, I began to notice other things about my old textbooks. Modern books are bigger—much bigger—than older ones. Observation of many bright summer interns tells me that this is not because students used to be smarter. I suspect that current technology creates the opportunity for including lots of fancy and informative graphics at moderate cost. There's also the fear of leaving out something important. To see this at work, just try plotting the number of pages of some well-known text as a function of number of new editions of that text.

A more important change is that lots of topics that were once considered "advanced" have migrated into "elementary" texts. While I believe that we really do learn everything important when we're very, very young, I doubt that this is the reason for the change. As time goes on, and new subjects are understood better, they simply seem to become more, well, simple.

The most important thing I noticed is that, in some sense at least, chemistry really has become physics, and so have some branches of biology. To put it more accurately, subjects that used to be thought of as separate and distinct are merging. And the place where they meet is in computation. All three subjects—physics, chemistry, and biology—are now deeply dependent on computation as their principal research tool, as are many other research subjects. The theme articles in this issue of CiSE contain several very good illustrations of this fact for the case of computational chemistry. Other theme issues have given other illustrations. Computational science has become the universal language of all other sciences.

But what is computational science itself? Naturally, I can't answer this question. However, thinking about it and trying to formulate at least part of the answer is a good task for everyone in the field. In my opinion, the definition of computational science is tied up with the description of how to train someone to be a computational scientist. If we know what constitutes education in a subject, we must have at least some idea of what the subject is. Many universities are now starting programs in computational science. CiSE will certainly be one forum for discussion of these programs, and some of the regular columns appearing in the magazine should serve as examples of materials for training computational scientists. Maybe CiSE should hang out a shingle saying, "Computational Science (and Engineering!) 'R' Us!"

New Editorial Member

William J. Feiereisen is the division leader of computer and computational sciences at Los Alamos National Laboratory. He has an MS and PhD, both in mechanical engineering and both from Stanford University. He is a member of the IEEE, IEEE Computer Society, ACM, and the American Institute of Aeronautics and Astronautics. Contact him at

65 ms
(Ver 3.x)