Congratulations to Cleve Moler, who has won the inaugural ICIAM Industry Prize 2023 for “outstanding contributions to the development of mathematical and computational tools and methods for the solution of science and engineering problems and his invention of MATLAB”.
I first saw Cleve demonstrate the original Fortran version of MATLAB on an IBM PC at the Gatlinburg meeting at the University of Waterloo in 1984. The commercial version of MATLAB was released soon after, and it has been my main programming environment ever since.
MATLAB succeeded for a number of reasons, some of which Dennis Sherwood and I describe in one of the creativity stories in our recent book How to Be Creative: A Practical Guide for the Mathematical Sciences. But there is one reason that is rarely mentioned.
From the beginning, MATLAB supported complex arithmetic—indeed, the basic data type has always been a complex matrix. The original 1980 MATLAB Users’ Guide says
MATLAB works with essentially only one kind of object, a rectangular matrix with complex elements. If the imaginary parts of the elements are all zero, they are not printed, but they still occupy storage.
By contrast, early competing packages usually supported only real arithmetic (see my 1989 SIAM News article Matrix Computations on a PC for a comparison of PC-MATLAB and GAUSS). Cleve understood the fundamental need to compute in the complex plane in real life problems, as opposed to textbook examples, and he appreciated how tedious it is to program with real and imaginary parts stored in separate arrays. The storing of zero imaginary parts of real numbers was a small price to pay for the convenience. Of course, the commercial version of MATLAB was optimized not to store the imaginary part of reals. Control engineers—a group who were early adopters of MATLAB—appreciated the MATLAB approach, because the stability of control systems depends on eigenvalues, which are in general complex.
Another wise choice was that MATLAB allows the imaginary unit to be written as
j, thus keeping mathematicians and electrical engineers happy!
Here is Cleve demonstrating MATLAB in October 2000: