Undergraduate embedded system education at Carnegie Mellon

Reviewer: Bayard Kohlhepp

Embedded systems courses have grown organically and abundantly at Carnegie Mellon for 30 years, and now cover a broad spectrum of topics. The authors mine this catalog to try to find order and reason, and to draw conclusions for future course planning. The course taxonomy they use to structure their reporting is valuable in itself, even if you don't read the details. Embedded systems are split into 11 different categories, spanned by six different crosscutting skills. For anyone describing an embedded system, building a formal ontology, creating project plans (such as a work breakdown chart), or even just hiring a new embedded systems programmer, this categorization is an effective way to describe the depth and breadth of embedded development. For each of the 17 categories mentioned above, the authors discuss several attributes: distinguishing features of the category, key skills and principles taught for that category, the teaching methods used to best engage students, and challenges particular to that category. These are the details you are likely to skip as you look for conclusions (unless you are an educator looking for advice on creating embedded systems courses, in which case you'll love this section). Some of the paper's conclusions are not surprising: students like outside speakers, small groups require less teaching overhead than large groups, and knowledge retention is better when working with actual implementations. Other conclusions might be very surprising; for instance, they find that experience with nonembedded systems frequently has to be unlearned in order to succeed at embedded design. And, in the end, it's decided that organic growth through informal networking might just produce better courses and topical coverage than centrally planned growth. I love happy endings! Online Computing Reviews Service