9: Looking back at CS for All

This week we zoom out.

Weekly update

This video is arriving late, so its introduction to this week's readings is retrospective. It responds to recent discussion reports on Turkle and Papert, diSessa, Lewis et al., Vogel, and Proctor et al., and then introduces this week's readings. The Week 10 video covers logistics for catching up and the rest of the course.

Summary

• One correct way to think? Schools may teach a single way of thinking less by intention than by working backwards from standards and assessments. In high school English, for example, the standards follow the New Critics in locating meaning inside the text alone. Naming that as one theoretical stance among many helps students whose questions about authors, contexts, and readers are just as legitimate.
• Concrete and formal (Turkle and Papert). Starting concretely, by seeing something work before abstracting, is a sound teaching approach and the one Making With Code takes. But Turkle and Papert would not call it a necessary developmental sequence, as Piaget did; some people work from the abstract to the concrete. And if, with Papert and diSessa, we think with tools and representations, even abstract thinking is concrete work with stuff.
• Computer literacy vs. computational literacy (diSessa). diSessa is not talking about knowing how to use a computer, and AI literacy proposals focused on how to use AI tools (such as prompt engineering) repeat the same mistake: they ignore what is going on behind the scenes and the ethical issues AI raises. A literacy framing is descriptive rather than prescriptive. Computational literacies emerge in context and should look different in different communities, more like music than like a single universal standard. Much of what we imagine computer science could be came from a narrow slice of Silicon Valley.
• Rethinking diSessa's three pillars. The social pillar can be split into two scales: the situated (participation, identity, and subject positions among the people around you, where individuals have real power to shape things) and the critical (the structure of society, where they have less). Media is not a third pillar alongside these, but an infrastructure underneath them that shapes practice at every scale, as algebraic notation transformed mathematics and social media transformed friendship and politics. Scholarship can move in two directions: designing new infrastructure for new practices, or critically analyzing how existing infrastructure shapes practice. Stories of AI harming people show how dependent we are on the technologies around us, and call for a relation of care with them, like the care good schools extend to the young people in them.
• Who defines computational literacy? Rather than institutions defining it from above, communities should be supported in taking responsibility for what computing looks like in their children's lives and schools.
• Stereotypes (Lewis et al.). The harm of a stereotype, accurate or not, is in the subject positions it makes available. In industry, for example, kinds of computing work can become racialized and assigned to geographies, such as testing work assigned to remote colleagues in India while new ideas came from the US.
• Language (Vogel). Accommodating linguistic diversity is not just translation, and programming languages are not like human languages: letting a Java course satisfy a foreign language requirement is absurd.
• Who decides what CS students learn (Proctor et al.)? Subject expertise does not imply pedagogical expertise; leading CS educators warn against letting computer scientists into the room too soon. Designing UB's advanced certificate in CS education meant choosing a middle ground between formal computer science and courses on pedagogy and conceptual frameworks.
• Are technologies neutral? This question keeps coming up across groups and weeks (see Winner's "Do Artifacts Have Politics?"). Society has been working it out for decades: internet companies have fought to be regulated as platforms rather than publishers, and the same question now faces AI companies' responsibility for what their models say.
• Fancsali et al. (2024). Twenty years of work built the infrastructure for CS in schools (standards, teacher certification, requirements), though much national support has since lost funding as attention turned to AI. New York State has been lackluster, but New York City invested more than anywhere in CS for All. This report from the research arm of that effort offers big-picture lessons for anyone shaping computing in a school or in policy.
• Philip and Garcia (2015). A look at the ideologies behind the first decade of CS for All, and at how well-funded nonprofits and corporations shaped how computer science entered schools. Related work treats literacy as a technology states use to socialize their populations, and argues that pushing computational thinking produces computationally tractable citizens who accept computational solutions to social problems, such as vaccine appointments available only through a website.

Readings

Fancsali, C., Mark, J., Hill, K., Li, X., Lee, J., Jain, R., & Flores, M. (2024 , December). Expanding Computer Science Education for All: Successes, Challenges, and Implications of NYC's CS4All Initiative. The Research Alliance for New York City Schools.

Philip, T. M. & Garcia, A. D. (2015). The schooling of Google: Ideology, Media, and the Corporate Shaping of Computer Science for All. Harvard Educational Review.