The Art of Computer Programming

Today's post is a review of Donald E. Knuth's The Art of Computer Programming. This venerable classic is all but forgotten perhaps in the fast-paced world of fly-by-night computing. I for one, purchased a copy of the 3-volume set about ten years ago while a professor, and I find it convenient as a reference, or simply for some good reading.

This page on mathematical induction is but one of many useful pointers for computer science students and practitioners alike.

In the case of computer science practitioners, referring to a page such as the one on mathematical induction, isn't something we do for fun, it's something we MUST do, for the benefit of our clients, and our ongoing growth. You are growing and learning constantly as a computer practitioner, or you're falling behind. There is no standing still.

The organization of the book is in three volumes: 

1. Fundamental Algorithms
2. Semi numerical Algorithms
3. Sorting and Searching

Want to know about garbage collection, before it was "cool"? The Art has it. Section 2.3.5. Thank you. Follow the symbols to understand: they are made to be language-independent so that any programmer can understand them without the baggage of a specific implementation. Perhaps, you'd like to make sense of the statistics you gather on a daily basis from your A/B testing? Check out Statistical Tests in Volume 2. I learned a lot about the statistical analysis package SAS as a doctoral student, then proceeded to publish tests to help prove that method B was better than method A with a certain confidence. SAS is great if you have access to it, but most of the time we perform simple enough tests that we could write a program to analyze the logs ourselves, and have the benefit of customizing the result presentation, given a little know-how gleaned from a good numerical analysis book. That's where Knuth's book comes in. Finally, if you want to know all about commonly-used searching and sorting techniques, and some not-so-commonly-used ones (but very useful), check out volume 3. Now, there are some sections dealing with tape merging, but keeping an open mind here, yesterday's big data was a lot smaller than today's, but the methods can be transferred.

The three-volume collection of computer science know-how!

In short, don't pass up the opportunity to purchase this classic and actually use it. Reading it cover-to-cover is useful for those who like to do so, but it is so easy to locate specific topics, it makes a great topical reference for the modern computer scientist.