A First Course in Scientific Computing: Symbolic, Graphic, and Numeric Modeling Using Maple, Java, Mathematica, and Fortran90

Landau takes a refreshingly different approach to teaching students scientific computation. The field can be considered as two parts. One, the older and more heavily used, is about the "traditional" numerical analysis. You crunch numbers, and you get numbers out. The other approach is symbolic algebra. Usually a text only deals with one type. Here, he teaches both. Plus, for each type, he offers the choice of two languages. For the numerical analysis, there is Fortran, version 90, and Java. While the symbolic algebra is performed using Mathematica or Maple. Ecumenical indeed! These are excellent choices of languages. Fortran still dominates legacy numerical analysis, with massive libraries of subroutines that one has to work with or maintain. While Java lets the student learn good object oriented practices. And Mathematica and Maple are perhaps the most common symbolic packages available.