An ability of an image to communicate the important features of a large data set makes graphical display of results an essential part of many computational research fields. Since I want to distribute my graphical results on the Internet, I chose a Java graphics package.
I am not an expert on graphics software, but let me list my ad hoc requirements for a Java graphics system. The software needs to display multidimensional arrays as grey-scale or color rasters. Simple graphics instructions draw lines and other elementary geometric figures. The package would include simple interactivity such as the zooming of image details and the display of image sequences as movies. Furthermore, the graphics software should allow custom plug-in processes for the display needs of particular images. A sophisticated volume renderer would offer interactive exploration of large, high-dimensional image cubes (similar to Rickmovie). Besides the data display, the graphics package would allow a flexible annotation of the images with titles, axes, and labels. A set of simple classes allows the authors to build simple interactive graphical user interfaces to their applications. These various goals could be accomplished by a combination of tools, but a consistent style of methods and parameters would make a single package desirable.
Since I lacked a suitable graphics package, I decided to use ImageJ Rasband (1998) of the NIH (National Institute of Health), rather than implement my own. NIH distributes a very successful non-Java graphics software and I trust the Java version will be further maintained and improved in the future. I decided for ImageJ since it offered a set of image processing routines and anticipated user extensions. Using a third-party software, of course, required that I compromise a few of my lofty goals.