The Mines Java Toolkit is open-source software for scientific computing. Why Java? Why not C++ (or FORTRAN or MATLAB)? My answer is in this paper:

The Java and C++ platforms for scientific computing

Scientific computing today requires more than simply processing arrays of numbers. Simulations of wave propagation and fluid flow on unstructured grids, free-form surface modeling, and interactive 3-D graphics are just a few examples of applications that exploit more complex data structures. Speed of development, extensive standard software packages, and cross-platform portability lead us to favor the programming language Java for today's scientific computing.

Hale, D., 2006, Center for Wave Phenomena, Report CWP–547.

[Report]

Think Java is too slow for array processing? See the results of our benchmarks comparing Java and C++ for digital signal processing. You might also review the results of the SciMark 2.0 benchmark provided by the National Institute of Standards and Technology.

The research highlighted here was performed using software packages written almost entirely in Java.