Customer Reviews

Python Scripting for Computational Science (Texts in Computational Science and Engineering)

5 star:		(7)
4 star:		(2)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

The author has 2 main goals:

1) To improve the productivity of scientists familiar with specific software systems (especially Matlab, Maple, and Mathematica) by teaching them to "glue" applications together.

2) To advocate Python as the preferred "glue" language. In his own words, "I hope to convince computational scientists having experience with Perl that Python is a preferable alternative, especially for large long-term projects."

He has certainly done a creditable job. As an expert in computational differential equations, he neglects neither efficiency nor correctness, while stressing both simplicity and reliability. In this sense, he has done a great service to the Python community.

The question is: What justifies the purchase of his book?

The answer is: Chapters 4, 9, and 10.

Contents:

1. Introduction--26pp

Very convincing arguments.

2. Getting Started With Python Scripting--38pp

Interesting examples.

3. Basic Python--56pp

A too-quick tutorial. Go to python dot org instead.

4. Numerical Computing in Python--48pp

Stellar explanations of vectorized array operations.

5. Combining Python with Fortran, C, and C++--36pp

Details use of Fortran2Py and SWIG. Mentions many alternatives.

6. Introduction to GUI Programming--70pp

Useful examples of Tkinter/pmw widgets.

7. Web Interfaces and CGI Programming--24pp

Good source of ideas.

8. Advanced Python--132pp

Deep and extensive. Includes: option parsing, regular expressions, data persistence and compression, object-oriented programming, exceptions, generic programming, efficiency.

9. Fortran Programming with NumPy Arrays--32pp

All about efficiency and re-use.

10. C and C++ Programming with NumPy Arrays--40pp

More about efficiency. NumPy C API, C++ objects, and SCXX.

11. More Advanced GUI Programming--73pp

Tedious discussion of both Web and standalone GUIs. BLT, canvas, cgi.

12. Tools and Examples--70pp

Excellent examples of PDE solvers, with a powerful GUI, but quite long and tedious.

A. Setting up the Required Software Environment--16pp

Wonderfully specific installation instructions!

B. Elements of Software Engineering--50pp

Python's strength! Very practical advice on modularity, documentation, coding style, regression-testing, version-control.

Strengths:

+ Downloadable py4cs package, esp. numpytools module

+ Great advice everywhere, e.g. CGI checklist, Pythonic programming, and trouble-shooting.

+ Concrete evidence for most assertions.

+ Very attractive presentation. Sturdy, high-quality cover, binding and pages. Brief, elegant code fragments (except in Chapter 12). Readable prose. No wasted space.

+ Available as 5MB pdf file, after purchase of hardcopy. Very nice.

+ Slides, installation instructions, and errata also at web site. Very professional.

My peeves:

- Not enough tables to be a useful manual.

- On p.428(#7) he points out that handling a raised exception is very slow. However, when I time his example with a positive argument, the try-except version is 20% faster (b/c the if clause is skipped), so he is actually giving bad advice for the general case. Luckily, he contradicts himself later, on page 685: "Exceptions should be used instead of if-else tests." The best advice: Avoid common exceptions in inner loops.

- The 10-page index is not as great as it at first seems. (See Martelli's Python in a Nutshell for a better one.)

- Pure interface functions should 'raise NotImplementedError', rather than 'return'.

- Exceptions should never be trapped mindlessly with 'except:'. That would hide your own SyntaxErrors!

- Too many exercises. (It's published as a textbook.) Since there are no answers, the exercises are useless for non-students. (See Lutz's Learning Python for effective exercises with answers.)

Overall rating:

This contains the best information on numerical programming in Python that I've seen. Though expensive, it could easily be your only Python book, given the excellent online documenation already available.

Python Scripting for Computational Science is both an introduction to the Python language and an excellent reference for the intermediate developer. The approach taken by the author is to present the language in the form of tasks to be solved accompanied by example code. As expected for a book on scientific computing the modules covered in the examples emphasize numerical packages but this in no way detracts from the applicability to general Python enthusiast.

What really makes this book more than just another Python introduction is that the author bridges the gap between complied and interpreted code. He demonstrates how the speed of execution of compiled code can be tied to the rapid pace at which scripts can be developed. Examples are provided for interfacing C, C++ and FORTRAN code with Python. Calls to precompiled applications are also covered and the examples were easily adapted to my favorite computational tools. One of the risks with doing numerical work in a scripting language is the possibility of straying into computationally intensive tasks to which interpreted code is not well suited . Latter chapters discuss how to identify these portions of your code and how to migrating these tasks to a compiled language.

I bought this book as an experienced programmer and Unix user expecting more of a "Numerical Recepies in Python" emphasis on the efficient implementation of algorithms which happen to be in Python. I should have paid more attention to the description.

This book is really more of a "Grad Student's Guide to Everyday Python Usage". I imagine it would be very valuable to a mathematics Grad student without too much programming or shell experience, looking for an alternative to Matlab. However, there is very little "Computational Science" in this book. Do NOT expect a cookbook of high performance algorithm implementations.

The book is a very verbose 700+ pages, all in an unexciting academic LaTeX format. The author works through idiom after idiom for accomplishing different tasks in fairly stand-alone sub-sections without much of a feeling of conceptual "flow" between them. It sort of feels like reading through the author's personal lab notes that he took everytime he learned a new language feature or trick.

If you are an experienced programmer, you will quickly get impatient with the verbose presentation that emphasizes idioms and examples instead of fundamental concepts and syntax reference tables. But, if you are an experienced programmer, you are not the target audience for this book.

Braddock Gaskill

Langtangen's emphasis here is on a reader who comes from a strong background in engineering or science, and is familiar with common computational ideas and has done some programming, but not necessarily in Python. The typical book on Python is aimed at a general programming reader, and the examples in such a book usually are quite elementary, from a computational viewpoint.

The merit of Langtangen's book is that he gets into a lot of computational ideas. This is not a trivial book. Aspects like parsing data in files, connecting to local and remote hosts, and interacting with programs written in other languages are covered. For the latter, the important cases of Fortran and C programs are explained. The choices of these languages is deliberate. In science and engineering, they are the dominant languages for raw computation. And you are likely to have legacy code written in these, that you cannot abandon while using Python.

I'm giving this book five stars because it was basically written for me. I don't mean that literally, of course. I say that because the usual methods of googling for answers and reading the manual do not work when you are trying to push the limits of what a tool is capable of doing. I do numerical computations for a variety of things -- finding patterns in large data sets, automating data collection and analysis, converting raw serial output into convenient CSV, plotting multidimensional datasets etc. Over the years, I have collected a large number of productivity habits with Matlab, which allows me to do ridiculously convoluted things in a short period of time. You just have to read the introduction of any Python manual to understand why I am switching from Matlab to Python. The problem is -- what will replace all these productivity habits? They need to be replaced with "Pythonic" habits, something that can take years of practice.

The beauty about Langtangen's book is that it runs through every one of those techniques. Instead of giving a basic example (what your google search would have provided) or a complete list of, ahem, useless techniques (what the manual would have provided), you get exactly what a seasoned data analyst needs to know to get moving with state-of-the-art commands. The author also discusses optimizations and alternatives in each chapter.

The book is also the best source for explaining *why* NumPy should be used by people working with large datasets. Folks love to create toolkits for Python, but some of these are a list of non-intuitive shortcuts that don't provide a substantial improvement over basic Python. Langtangen goes through the pain of explaining the benefits of the package (chapter 4.1.4), so that you can decide for yourself if NumPy is useful for your application.

I will not comment on the parts of the book that deal with C and FORTRAN integration because I leave that to more able programmers. I also will not comment on the extensive GUI building chapters because I do not build GUIs. I will point out, though, that I have derived full value out of this book simply by reading, and re-reading chapters 2, 3, 4 and 8. Some will argue that there is too much "basic Python" in these chapters for the whole to be considered advanced computational science -- my opinion is that even when the author describes "basic Python", his examples and intuition make it so that even one who has read a couple of reference books cover-to-cover will learn something about using "basic Python" to perform numerical analysis in a more efficient way. In fact, the book is a testament to doing really convoluted things in a really compact and elegant manner!

When I first got ahold of this book I had just finished learning all the gory details of good numerical codes. But when developing tests for simple cases I found that development went way too slow, so someone suggested I learn Python. This book provides a great demonstration of how python can supplement your existing codes. Either by organizing the tests, formatting output, or just adding pretty interfaces.

This book contains a lot of the necessary extras that a scientist or engineer must do to get his work going or finished, which is too pedantic to be taught in most courses. It shows the power of Python over some other scripting languages for this purpose. It is definitely one of the best references on my book shelf.

I've bought what seems to (my wife) be every Python book out there and I can't tell you how sick I am of spam, spam, spam code! (trivial and obfuscated Python code examples with a common theme focused around one Monty Python skit or another...) Spam code seems to prevail in other Python books.

Here finally is a book with code examples that are very clear, are immediately useful to the serious programmer and filled with real life discourse on relative performance differences between Python and other languages that have a reputation for speed. There are clear examples of 'number crunching', producing images and even video animations, hooks into other scientific packages such as MathLab, etc.

If you are interested in really learning Python, want to come away from an hour or twos worth of coding experience with a module or two that you can use tomorrow and are not interested in code examples extolling Monty Python silliness, then this is the book for you.

While this book is about twice as expensive as many of my other Python books, I wish I had purchased this one first. Even though I've been using Python, seemingly every day, for two years, I kept finding nuggets in this book with what seemed to be every turn of the page. My focus right now is processing extremely large data sets of binary data but I'll soon be looking at image processing and I know I'll be reaching for this book over and over again. Don't hesitate! Just buy the book!

If you want to learn Python, you should get it. Author do not build some "big" application (like "internet store software" or "bookstore software") from beginning to end, but rather give you a lot of practical examples of using python.

It is not like in others book that examples include only learned functions/methods, but use topics from the rest of book (you have example on page 25 and note that explanation of this and that function you found on page 543). By that you have interesting examples to use in real-world problems, not only examples to explain freshly learned topic.
In other books interesting examples of use python you found on page 3234, because only when author introduce all useful functions. In this book nice examples is even on first pages.

You learn how to use numerical packages (numpy) in python, using some useful tricks on lists and arrays, introduce to using graphical interface in Tk.

I bought this book, just for a couple of the chapters, but i found myself using more of this book then i expected, and reading all the chapters(even the fortran stuff). I found this book better then all my other "scientific python" books, in that my other books really built toyish apps. This book is meant for people doing production computational science work in python. It doesn't have much btw on super computer's programming and python, aside from a lot on how to integrate c/c++/fortan libraries(i.e. anyone doing major work in python probably is integrating to things like Tesla/Hadoop/mpi ... etc ... and the book didn't go to that level).

As an intermediate Python programmer, this excellent book has become my go to reference for useful intermediate and advanced techniques that I can locate and learn quickly. The writing is clear and not overly verbose. In addition to a wide array of numerical and scientific examples, the book is helpful for a wide range of programming issues, such as gluing together disparate legacy applications, interfacing to C++, regression testing numerical code, building GUI's, web programming, etc.