An Introduction to Imperative Programming

To this end, the course is accompanied by a series of programming projects in all languages covered (MIPS, C, Java) that increase in size and difficulty.

One part of learning the practice of programming is also to learn about data structures and algorithms that solve interesting problems. We try to provide this experience through non-trivial programming projects that often rely on clever data structures and algorithms that make the project feasible in the first place (shortest paths in route planning, acceleration structures in ray tracing, dynamic programming, etc.).

I consider it important that students are able to identify the different levels of abstractions that are at work in programming and they can mentally move between these abstractions without problems. Seeing a C/C++ program and actually having a vague idea of what happens on the machine is not only helpful to eliminate “cargo cult” thinking but also vital when it comes to performance or security considerations. Therefore, the course intentionally starts with computer arithmetic and machine code programming to convey an impression on how programs execute at the lowest level. This enables an understanding of the abstractions higher-level imperative languages provide to make programming machines more bearable and scalable, what these abstractions actually abstract from, and what the cost of these abstractions are.

Being able to understand how computers execute is one thing. It is equally important to understand how to describe programs and their execution in mathematical terms to be able to talk about concepts like termination, undefined behavior, correctness, etc. in a meaningful way. It is important to understand that the details of these notions are so intricate, subtle, and important that it is impossible to leave it at hand waving. Additionally, one can make the experience that the process of formalizing things uncovers and clarifies ambiguities and makes precise what one is actually talking about.