Computational Physics: A Physics-First Approach
Publisher
Haverford College Libraries
Haverford Libraries Support
OER Grant
Publication Date
8-4-2026
Description
This resource is designed for upper-level undergraduate computational physics courses. It includes Jupyter notebooks for eleven modules, each built around a physical system rather than a numerical method. Every module opens with a physical system, develops the physics far enough that a computational obstacle becomes unavoidable, and only then introduces the method that clears it. Root-finding arrives because the logistic map's period-doubling bifurcation has no closed form. Monte Carlo integration arrives because the partition function of a magnet has 2^N terms. The method is the answer to a question the physics already asked.
Two consequences shape the notebooks:
Results are verified, not asserted. Where theory predicts a convergence order, a scaling exponent, or an equilibrium value, the notebook computes it and prints the comparison. Students see Simpson's rule convergence order: 4.00 (expect 4), not a claim that it should be 4.
Modules build on each other. Module 1 ends with an open question about step size that Module 2 answers formally and Module 2's second lecture revisits at the level of floating-point round-off. The Standard Error of the Mean derived in Module 8 reappears as the Monte Carlo integration error in Module 9b.
Disciplines
Physics
Keywords
computational physics
Comments
A volume in the Haverford College Libraries' Open Educational Resources series.