INFRASTRUCTURE BUILDING

We have been criticized for spending a lot of energy building our infrastructure. By this I mean building software packages, SEPlib, vplot, SEP3D, ratfor90, our reproducible research makefile rules, and Sergey's optimization library. Unlike SU, which was developed at CSM, our software has hardly defined industry standards, nor is it widely used outside SEP (except for Joe Dellinger's vplot which seems to show up in nearly every issue of Geophysics). We are not here to produce widely used software, though that is gratifying when it occurs (Rick Ottolini's movie program is another example). We build our infrastructure to get our work done, and to pass along reusable results of our older generation to the next. We do use some commercial packages, and despite hefty discounts, I don't think we get our money's worth from them. I wish we were paying on a ``per use'' basis. Although some powerful past personages here at SEP have done heroic 3-D work with little more than our SEPlib, vplot, and a bare-bones Fortran90 compiler, propagating those skills to our broader population of students requires some infrastructure. I'm very happy (absolutely thrilled) that Biondo Biondi and Bob Clapp have built SEP3D for us.