Readings from Thanksgiving Break

I’ve had a few days off of work and that’s given me a bit of a chance to do some general computer science education blog reading. The top posts I’ve read, in no particular order:

  • Teasing apart the issues of women in computing: the impact of Hollywood – TL;DR version is that the NYTimes has recognized that media does not depict women in engineering or computer science roles (as have many other sources) and suggests that Hollywood exposure of women in CS would solve the real-world lack of women in the field.
  • Guzdial disagrees with the NYTimes call for a Hollywood movie about CS-ers and raises a series of important research thoughts and questions in response. He claims that the biggest challenge in CS Ed is getting students and teachers to even consider computer science. I whole-heartedly agree. Based on my experience with Software Carpentry – data scientists do not want to invest the time to learn to program until they realize exactly how much time it will save them. [That said, I’d really love to run a study somewhere that evaluates exactly how much time scientists save themselves by doing scientific computing with programming, version control, and all that jazz.]
  • Why arts and social science need code: testimonials – I seriously need to start collecting examples of people who use code in nonstandard/unexpected ways. These examples tend to be extremely motivating for students that I encounter at the K-12 level. 
  • The Audrey Test: Or, what should every techie know about education? – I think Watters encapsulates a major let’s-just-build-it problem that happens in CS education. Programmers tend to love hacking. So, of course, when faced with the problem of lacking diversity in CS, they want to build new tools to fix the issue. They want to use what they are good at – code – to change the playing field. I understand. I rely on my own skills when tackling new problems as well. But it’s important to orient ourselves in the problem before tackling it. Otherwise, we just build and rebuild similar products
  • In a different vein entirely, I came across this blog post about scientific computing homogenizing in Python. It argues the benefits of migrating all your data science work into Python (rather than a mashup of Python, R, MATLAB, Ruby, etc). The sentiment of the author near the end resonated with me: Rather than considering What’s the best tool for the job that I’m willing to learn and/or tolerate using? he asks himself Is there really no way to do this in Python