Guiding principles - NormandaleWells/CSn GitHub Wiki

This work is guided by a few basic principles:

Working code

I want to eventually supply working code for everything discussed here, in at least these languages:

• Java is a nice language for the CSn sequence, since we don't have to worry (much) about memory management. It also has a simple syntax, and is easy to learn.
• C++ is my main language, but until C++11, I would never have considered using it for this course. Modern C++ has features that make memory management nearly trivial, and it has no close peer in regard to generic programming and efficiency.
• Python has become a very popular language, largely due to its ability to interface with libraries like NumPy, SciPy, and SymPy.

There are also a couple of new languages which I find intriguing, and which may get some support here:

• Kotlin "fixes" much of what I don't like about Java.
• Rust retains the uberefficiency of C++, but has interesting facilities to control object access and lifetime.

Standard libraries

One of the ironies of teaching a data structures and algorithms course is that every language's standard library already contains the data structures and algorithms studied in that course. I've chosen here to ignore that fact. The idea here is that the purpose of the course is to show the student how these algorithms work, and the best way to do that is to provide an implementation she can study and modify.

But for heaven's sake, in real-world code use the library that comes with your language. It is probably more efficient and certainly better tested than anything you'll find in the code repository here.

Array indexing

Every modern language starts arrays at 0, and most books on algorithms completely ignore this simple fact. (Sedgewick and Wayne's Algorithms, 4th edition is one notable exception. See references.) I often wonder how many off-by-one errors are caused by programmers improperly translating a book's 1-based array pseudocode into their 0-based language of choice.

Furthermore, 0-based array indexing invites the use of half-open ranges, which are used extensively throughout the code here. For good examples of how the use of half-open simplify programming and make it easier to analyze code for correctness, see the code for upper_bound, lower_bound, and quicksort_3way.

Writing style

You can have my Oxford comma when you can pry it out of my cold, dying hands. After all, without the Oxford comma, I could claim that some of my major influences would be my parents, Robert Sedgewick and David Gries.