Sub Ubi

By and large, I’m unimpressed with the quality of literature research that takes place in most papers in the area of CS education.

The papers I posted earlier are largely from Google searches and the ACM Digital Library. Neither was there any real search of the IEEE in my process, nor any stack crawling. In short, I looked up all the papers I could quickly and easily search and download, because I wanted information fast.

I found a great deal of interesting stuff regarding pedagogical programming languages, pedagogical programming environments, the quality of compiler error messages and the like. As I read through these papers from the 60’s, 70’s, and 80’s, I wonder why more of them haven’t been cited by papers written in the late 90’s and early 00’s? Given how easy it is to search on these topics, and how easy it is to download and print papers related to computing, why can’t authors at least do a half-assed literature review, instead of one where they fail to mention work that was exactly like theirs, published in the same forum, just 6 years previous?

Every new language that hits the scene (be it C, C++, Java, C#, …) has caused a shift in the first course in computing (CS1, as it is sometimes called). Pascal gave way to C and C++, which yielded (in a remarkably short time) to Java, and now people are looking at C# for a first course in OOP; speak to me not of the issue of “programming paradigm” even! Sadly, with every shift, we loose the tools and experience that scaffold the novice programmer. Years ago, there were some excellent tools for supporting students programming in Pascal; oops! Gone! Then, we built tools to make programming in C a better experience, interpreters and debuggers to support novices as they explored the language. *POOF* Gone.

The latest generation of tools don’t always provide adequate cover for the lessons learned of old. It’s as if every language is different, and every generation knows best; CS-ED really needs to get it’s act together. I’m glad industry is made up of brilliant professionals who can switch from one tool to another every 6 to 8 years, but it takes time to develop, test, and revise educational tools, not to mention materials to support learning, rubrics, and everything else that is necessary for the educational enterprise.

Later, or tomorrow, I’ll write about a conversation I had yesterday that does a nice job of coming at some of these issues from a more disciplined perspective, and provides some interesting food for thought, independent of whether or not everyone has done an adequate review of the literature.