As more and more of our exploration is in knowledge spaces like this, the divisions between disciplines are going to be blurred. We have talked a little bit about specialization and generalization. It certainly is true that there is a long tradition of both interdisciplinary work and of scientists being generalists. The relative importance of interdisciplinary work has waxed and waned. It seems to me that after World War II there was a tremendous move toward specialization in science, which began with big federal funding of research. Perhaps this was the beginning of the various pressures that we have talked about in terms of being in the grant rat race.
I believe that information technology will start to reverse this trend; in fact, it already has started to make it possible for individuals to find out about things more efficiently, for individuals to become multiple experts. Indeed, because of the pace of change and the various tasks that people are going to do during their lifetimes, it will be almost mandatory for people to become multiple experts.
I certainly don’t mean that interacting with people over the Web should replace mentoring. As collaboration tools get better and better, they are not a way for people to avoid communicating with each other, or avoid mentoring each other—they are other ways for people to communicate and mentor. I can tell you the innermost contents of my mind as well over the Internet as I can in person. In fact, I may be even less inhibited about doing so.
I would like now to show you what I am talking about in terms of the Biology Workbench. Figure 5.1 shows what we call the National Computational Science Alliance (Alliance) Information Workbench, for which the Biology Workbench is the prototype. The user interface in this diagram is the Web browser. The Web browser connects to the guts of the Workbench, which is the Workbench server. The Workbench server translates formats, creates queries that databases can understand, and drives application programs. The application programs have various interfaces, as can the information sources, and can be written in different languages. All of these can be readily tied together by a powerful scripting language such as Practical Extraction and Report Language (PERL), which is the easiest to use. But you can script in C as well as in PERL.
The key is that we now have the ability to take, for example, a whole series of databases in varying formats, such as various molecular biology databases, and a whole series of application programs, such as programs to visualize molecules, align molecular sequences, and construct phylogenetic trees of relationships based on those sequences, and so forth, and make them all look like one program. You interface through them and point and click as though you were on a Macintosh.
I know that this is not a Macintosh, but the point is to make everything in the world look like one. This is what we are really about. I tend to think in terms of biology, but basically the challenge is the same for chemistry. In chemistry you have databases that give you physical properties of chemicals; you have databases that give you structures, and so forth. For each discipline, what you really want to