I had a fun cohort of students in CS222 this semester. Each team designed a final project of the "connect to a Web service and do something with the data" variety. In many ways, their foibles and experienced were akin to past students, but something that stood out to me was how well they asked each other questions. It is not clear whether I have done something different to foster this; one possible factor is that I have improved the requirement that teams give a technical lesson as part of their presentations.
We cover a lot of ground in CS222, and so I am concerned that next time, I will have only fourteen weeks of instruction rather than fifteen. The university is changing the academic calendar as of Fall 2026, and it will be hard to find a week's worth of work to remove. Also, students coming in will have two fewer weeks of classroom experience programming; given that many students do not do much independent practice, this will have a significant effect on downstream courses. Coincidentally, I just got out of a meeting this morning where we talked about how much more we want to teach within the undergraduate major, and yet we will have eight fewer weeks in which to do what we are already struggling to fit in.
My course requires students to read and evaluate "project code," which is a term I define but that a majority of students don't understand. Either they don't read the instructions or they lack the ability to distinguish between code-for-learning and code-to-do-something. As a result, many of them look back on their own code that was just purely pedagogic; others find code on GitHub, which I recognize is a leap of faith, but they look at code that is just someone else's classroom or tutorial project. This leaps out to me upon inspection, but because they have only ever read pedagogic code, I don't think they recognize it. Hence, rather than just refine by definition, I think I need to point them toward specific examples of projects that are appropriate. I hate to do this because there is so much code out there and I want them to find projects that are interesting to them. Yet, despite this encouragement, almost no one has ever done that, so maybe that's a dream not worth chasing. I would need to set up some guardrails to prevent two students from evaluating the same block of code, which is also something I am not keen on. Perhaps I need to set up some kind of draft.
I have had a note in my planning spreadsheet for some time to remind students that a good way to think about SRP is responsibility to whom. Also, though, there's a second edition of Clean Code out now that I haven't read yet. I need to see if the new edition is worth switching to, especially around explanations of OO principles that often trip students.
My final note is about the final exam. Many years ago, it comprised only reflective questions, and those gave me an insight into the student experience. A few semesters ago, I was concerned about students blowing steam, and so I added a content question as well, to make sure students could express some of the fundamental class topics. One of these questions has been perennially challenging to articulate correctly. Essentially, the question is trying to see if students understand that they should be able to make a list of unit tests they have yet to write and that these should be SMART. The trouble is that if I remind them that each step has to be SMART, then I have given them the answer. What happens is that students will say that their first step is to "write some tests," for example. That is not wrong, but it also doesn't illustrate what I'm trying to get at. Relevantly, I think many student teams continue to struggle with this idea throughout the semester, that they should always have a goal to work toward. TDD is supposed to push them in this way, but in practice, many fall back on old habits of just plowing forward without discipline. So, I think this points toward two other action items for me: first, to use more class time to have them practice articulating what their immediate next steps are, like Kent Beck does in Test-Driven Development By Example, and also then to use a similar framing to ensure that they can do this individually, as an outcome of the class.
No comments:
Post a Comment