The Minds Behind Numerical Computing – 7 People Redefining Software
Seven pioneers helped establish numerical computing through electronic scientific machines, optimization, error analysis, matrix algorithms, data analysis, and MATLAB.
TL;DR
Numerical computing emerged when electronic computers, optimization, error analysis, matrix algorithms, statistics, and interactive mathematical software converged. Von Neumann and Goldstine helped establish scientific electronic computation; Dantzig created a defining optimization algorithm; Hamming and Wilkinson made computational error a disciplined subject; Tukey broadened computational data analysis; Moler made matrix computation interactive through MATLAB.[1][5][7]
Why you should read it anyway
Scientific computing changed mathematics from a discipline limited by hand calculation into one capable of simulating weather, structures, fluids, circuits, economies, and physical systems. The challenge was never simply “do more arithmetic.” It was deciding which algorithms remain trustworthy when every number is rounded and every computation has finite cost.
Imagine where Numerical Computing would be without them
Without this lineage, electronic computers would still have been used for science, but much more effort would go into bespoke code and unstable numerical methods. Optimization, matrix computation, simulation, signal processing, and engineering analysis would mature more slowly.
Time Estimate of how many years we would be hindered without them for human progress
Editorial counterfactual estimate: 8–15 years. Wartime and industrial demand ensured rapid progress, but the combination of computer architecture, numerical analysis, algorithm design, and reusable mathematical software materially compressed the timeline.
The 7 people behind Numerical Computing
1. John von Neumann
Why they matter: Von Neumann helped turn automatic electronic computing into a practical instrument for mathematics and science. At the Institute for Advanced Study he led the Electronic Computer Project and helped articulate the stored-program architecture that allowed one machine to carry out many numerical algorithms.[1] His influence extended directly into numerical methods for differential equations, Monte Carlo work, weather computation, and scientific simulation. The critical shift was conceptual: mathematics could be reorganized around machines capable of executing long, reproducible sequences of arithmetic operations.
2. Herman Goldstine
Why they matter: Goldstine served as the Army liaison to the ENIAC project, introduced von Neumann to the machine, and then worked with him at the Institute for Advanced Study.[2] He co-authored foundational reports on electronic computing and numerical methods and later led mathematical-sciences research at IBM. His contribution sits between hardware, mathematics, and institution building: numerical computing needed both machines and communities capable of developing reliable mathematical software for them.
3. George Dantzig
Why they matter: Dantzig created the simplex algorithm and became one of the founders of linear programming.[3] The algorithm turned large optimization problems into computations that electronic computers could repeatedly solve, influencing transportation, manufacturing, scheduling, logistics, and operations research. Numerical computing became economically transformative when machines could optimize decisions, not merely evaluate formulas.
4. Richard Hamming
Why they matter: Hamming worked at Los Alamos and Bell Labs across numerical analysis, digital filtering, coding theory, and computation. The National Academies notes his continued work on numerical integration and numerical analysis as electronic computation expanded.[4] His contribution was a culture of computational reliability: algorithms had to be understood in terms of error, approximation, stability, and the limitations of finite arithmetic.
5. John Tukey
Why they matter: Tukey connected statistics, computation, signal processing, and data analysis. His work helped popularize exploratory data analysis and, with James Cooley, the fast Fourier transform lineage that dramatically reduced the cost of spectral computation. In this article his role is broader than one algorithm: Tukey helped establish a computational style in which data could be transformed, visualized, and interrogated interactively rather than treated only through closed-form formulas.[6]
6. James Wilkinson
Why they matter: Wilkinson became one of the central figures in numerical analysis for electronic computers. MathWorks’ history of MATLAB identifies him as a leading authority on numerical linear algebra and traces modern matrix software back through his work.[5] His backward-error analysis changed how researchers judged algorithms: a useful numerical answer is often best understood as the exact answer to a nearby problem rather than as an imperfect approximation to the original one.
7. Cleve Moler
Why they matter: Moler created the first MATLAB as an interactive matrix calculator so students could use LINPACK and EISPACK without writing low-level Fortran.[7] His contribution was accessibility. Numerical linear algebra moved from specialist libraries into an interactive environment where engineers and scientists could express matrix calculations almost as they wrote them mathematically.
How they each differ from one another
Von Neumann and Goldstine represent the machine-and-method foundation; Dantzig optimization; Hamming reliability and applied numerical thinking; Tukey computational data analysis; Wilkinson rigorous numerical stability; Moler interactive mathematical software. The field formed because these layers reinforced one another.
Final Take
Numerical computing made approximation respectable by making it analyzable. Modern scientific software does not merely produce numbers—it carries assumptions about error, conditioning, convergence, and computational cost. That discipline is what allows simulation to stand beside experiment as a major tool of science.
Works Cited
- 01
- 02
- 03Stanford Engineering — George Dantzig engineering.stanford.edu
- 04National Academies — Richard W. Hamming Memorial Tribute nationalacademies.org
- 05MathWorks — The Origins of MATLAB mathworks.com
- 06
- 07MathWorks — A Brief History of MATLAB mathworks.com
CodeHistory is a living archive. Citations document the evidence used for this edition; later evidence may refine the account.
Submit a research lead