MATLAB is the principal numerical math system for matrix computation on platforms ranging from DOS PCs up to scientific superminicomputers. The big news in version 4.0 and higher of MATLAB is the addition, at last, of two features that have been needed for years. First, MATLAB has done an excellent job of implementing 3-D color graphics, making it the last major piece of scientific software to offer this capability. Second, The Math Works has finally introduced a proper debugging facility for M-files, the command scripts that are the basis of most MATLAB operations.
MATLAB’s language is like an interpreted version of C, so it’s relatively simple to implement a toolbox — a toolbox is just a set of M-file scripts and some new function definitions. Image processing, statistics, and electrical engineering toolboxes (control, signal processing, system identification) all appeared shortly after the first release of MATLAB. I looked at two new ones: the Symbolic Math Toolbox ($595) and the Neural Network Toolbox 2.0 ($895).
A Neural Network Toolbox is a natural for MATLAB, since every neural-net model is implemented as a matrix computation. Version 2.0 includes a whole megabyte of M-files (in highly compact format at that). The set covers backpropagation; perceptron learning; delta rule nets; Kohonen and Hebbian association; Hopfield recurrent networks; and nine detailed applications, including character recognition. The 400-page manual is one of the best textbooks on the subject I’ve seen yet.
The Symbolic Math Toolbox, instead of being yet another element in the endless repertoire of matrix computations, is a significant departure for MATLAB. Actually, it’s Maple, appearing as a new set of MATLB commands. The Math Works licensed Maple code from Waterloo Maple Software and packaged it as a standard symbolic-math toolbox and an $895 Extended Toolbox. The standard toolbox has about 60 Maple functions, including the linear algebra and ordinary differential equations sets, while the extended toolbox is essentially the full Maple package, including programming capabilities. The speed of the Maple functions inside MATLAB is impressive, especially since MATLAB just calls Maple through its MEX gateway, and MATLAB doesn't impose much overhead of its own, being a simple command-oriented system.
Seiter, Charles. (November 1994). MATLAB 4.1. Macworld. (pg. 81).
The MathWorks, Inc.
Applications
out of 10 (0 vote)
107 / 
2015-08-26 / 
01bd7fad56b16c84256d24c977d2d1fefdcd9b81 / 
/ 
MATLAB_4.2c.1.sit
Emulating this? It could probably run under: 
Basilisk II
