In 2026, I don't want to do numerical programming in C. That was fine 30 years ago, but today, I expect to have garbage collection or to be able to multiply a matrix as A×B.
I took Mark Newman’s course some years ago. It was fantastic! Geared at sophomore/ junior year physics major — someone who had completed the basic intro sequence. I am sure this book is also great.
I did a few courses across academic years that were based around this book and it's very handy skills to learn. Whilst perhaps not in the moment, it's a good introduction to implementing functions and equations, before you lead on to the next steps of specific functions and methods of analysis alongside hpc with parallelization.
There's actually a source tex file bundled with exercises with a custom setup.tex which makes me believe the whole thing is bespoke. Might be wrong though
I read most of the 1st edition (busy), I'm sure it hasn't changed much to the 2nd. I would say it's rather good at an introductory level to the subject!
It definitely targets physics undergrads who have never programmed so if that's not you then you may feel friction during some chapters. If, like me, you are much more developed in programming than physics you might just want to do the exercises in the first few chapters to check your knowledge and move on to the good bits.
If you're looking for something more rigorous I would bet [Numerical Recipes](https://numerical.recipes/) is better (I haven't read it but I want to; see "busy").
Looks like not much. The book is about using Python to implement numerical methods, mainly about teaching the Python part, and that's all explained. You might be missing motivation if you don't know any physics, but even so, basic mechanics using differential equations seems to be enough to give context, at least for the earlier parts
Just to give a bit of flavor, I was a math + physics major in the 80s. The physics curriculum had some oddly named courses such as "theoretical physics" that were not really physics courses but were meant to give you the math and computational background needed for the more advanced courses or for graduate work. The math was stuff that wasn't covered extensively enough in the math major courses, such as vector calculus.
I was a math + physics major in college, in the 80s. Thankfully, our differential equations course covered both analytical and numerical integration. We also took a course in the math department called "numerical analysis" that got further into it and also dealt with the foibles of floating point arithmetic.
I think the course by Richard Fitzpatrick is a much better selection of content if you want to actually do computational physics: https://farside.ph.utexas.edu/teaching/329/329.pdf
Should be modernized to Python or similar.
In 2026, I don't want to do numerical programming in C. That was fine 30 years ago, but today, I expect to have garbage collection or to be able to multiply a matrix as A×B.
If that's what you want, use Matlab. High-performance scientific computing is still using C, C++ +/- CUDA, or Fortran, with Rust a growing segment.
A bit surprised Sussman's and Wisdom's book hasn't yet been mentioned: https://mitpress.mit.edu/9780262028967/structure-and-interpr...
I took Mark Newman’s course some years ago. It was fantastic! Geared at sophomore/ junior year physics major — someone who had completed the basic intro sequence. I am sure this book is also great.
I did a few courses across academic years that were based around this book and it's very handy skills to learn. Whilst perhaps not in the moment, it's a good introduction to implementing functions and equations, before you lead on to the next steps of specific functions and methods of analysis alongside hpc with parallelization.
The matplotlib chapter seems fairly barebones but I remain in awe at this gorgeous latex work
Isn't it a pretty standard book/memoir template?
He could have invested in a Python syntax highlighter. I use minted, myself, but I'm sure there are many alternatives.
There's actually a source tex file bundled with exercises with a custom setup.tex which makes me believe the whole thing is bespoke. Might be wrong though
https://websites.umich.edu/~mejn/cp2/exercises.html
By the way, I use typst now, so I don't have to worry about highlighting anymore!
Could somebody provide some opinion on the book and/or accompanying course?
I read most of the 1st edition (busy), I'm sure it hasn't changed much to the 2nd. I would say it's rather good at an introductory level to the subject!
It definitely targets physics undergrads who have never programmed so if that's not you then you may feel friction during some chapters. If, like me, you are much more developed in programming than physics you might just want to do the exercises in the first few chapters to check your knowledge and move on to the good bits.
If you're looking for something more rigorous I would bet [Numerical Recipes](https://numerical.recipes/) is better (I haven't read it but I want to; see "busy").
No, Numerical Recipes isn't better. Or worse. It's a different book on a different topic, with there topic very clearly advertised in the title.
It's a series of... numerical recipes. Nice descriptions of many numerical algorithms sufficient to use them.
It's not focused on physics. It's also not rigorous.
The Sussman / Wisdom reference is rigorous.
Why would you post about a book you haven't read?
lol. lmao even.
What physics do I need to know to follow this book?
Looks like not much. The book is about using Python to implement numerical methods, mainly about teaching the Python part, and that's all explained. You might be missing motivation if you don't know any physics, but even so, basic mechanics using differential equations seems to be enough to give context, at least for the earlier parts
> Exercises by chapter
Click on a chapter to download:
Chapter 2: Python programming for physicists
Chapter 3: Graphics and visualization
Chapter 4: Accuracy and speed
Chapter 5: Integrals and derivatives
Chapter 6: Solution of linear and nonlinear equations
Chapter 7: Fourier transforms
Chapter 8: Ordinary differential equations
Chapter 9: Partial differential equations
Chapter 10: Random processes and Monte Carlo methods
Chapter 11: Data science
Just to give a bit of flavor, I was a math + physics major in the 80s. The physics curriculum had some oddly named courses such as "theoretical physics" that were not really physics courses but were meant to give you the math and computational background needed for the more advanced courses or for graduate work. The math was stuff that wasn't covered extensively enough in the math major courses, such as vector calculus.
Weber's Electrodynamics.
Only after working through Rudin’s Analysis first.
[flagged]
I was a math + physics major in college, in the 80s. Thankfully, our differential equations course covered both analytical and numerical integration. We also took a course in the math department called "numerical analysis" that got further into it and also dealt with the foibles of floating point arithmetic.
For us, it was all in FORTRAN.
[dead]
[flagged]
computation will revolutionize physics.
I hope that's sarcastic. Physics is the original computational science.