Thursday, November 19, 2026

Rubber Duck Debugging

I’ve spent a good bit of time, online and in-person, lately talking to people about rubber duck debugging. It was one of the topics of my talk at CSTA NY recently. There was a lot of discussion about it at the end of my talk and the ideas flew around. I thought it might be a good idea to share some of what I have learned.

The basic idea behind rubber duck debugging is that before students ask their teacher for help on a problem they first explain what is happening to the rubber duck. By explaining what is happening and what should be happening to the rubber duck the student will often realize what they have done incorrectly. Explaining problems like this to another person is a common debugging technique for even professional developers. I should say, especially for professional developers!

What usually happens is that students are asked to first explain things to the duck. If that fails to show the correct result than the student is to explain it to a peer. Usually, but not always, to a person sitting next to them. Only than can they bring the teacher into the problem. Saving teacher time is a side benefit not the real goal though. The goal is to help students solve their own problems by carefully working at understanding what is happening in their code.

Does it work? Well, a lot of teachers swear by it. It is common enough that a rubber duck seems to be the unofficial mascot of Computer Science education. They were everywhere at several CSTA events that I have attended.

Teachers introduce and use the ducks in several ways. One of the most common is to issue a duck to every student. Some teachers even have adoption (aDUCKtion) ceremonies and certificates. The idea is to create a sense of ownership or teamwork with the duck. Making sure that the duck doesn’t just disappear into a backpack never to be seen again can be tricky. Some teachers require that the duck sit next to the computer or keyboard at the start of class for example.

Other teachers assign a duck to each computer and it stays there from class to class and student to student. Still others have a “duck pond” and students go to the pond to borrow a duck. Finding a system that works depends a lot on the school, budget, and student culture.

Getting students to use the duck is the next issue. Teachers have to be firm in asking students if they have talked to the duck first. There can be incentives of a sort as well.

One teacher told me about having ducks that quack. If a student talks to the duck and that solves the problem they can announce the fact by having the duck quack. Of course keeping the duck quiet at other times can be a classroom management issue. Again, something that works or not depends a lot on the classroom environment. Regardless, providing positive feedback for student success can go a long way towards getting students to use the technique.

One middle school teacher reported that quacking ducks would never work in a middle school. In my limited middle school  teaching experience, I would tend to agree. This teacher uses duck stickers that are attached to student notebooks. The duck is not the magic, getting a student to talk out their problem and their code is the magic!

Getting students to read and analyze their code is a particularly important part of teaching computer science and programming these days. Arguably, in the day of AI written code, being able to read, understand, and explain code is more important than ever. If a rubber duck can help that is a good thing.

Saturday, November 14, 2026

CSTA New York Conference 2026

The state wide New York state CSTA chapter is running their first conference. We’re at the State University of New York New Paltz today. I missed the CSTA New England conference last weekend (other commitments) so itis great to be here. I am enjoying connecting with friends I have made over the years. CSTA is a real community.

First up, the keynote Philip Courtney CEO of Urban Arts. Urban Arts teaches video game design and emerging tech. Philip has an interesting background that started in the theatre and evolved into education. The idea behind Urban Arts is a blend of art and technology. I love the idea of involving the creativity of art combined with technology. I have always believed that computing was a creative endeavor. He brought  two alumni of the program who are now in university. These students showed what they did at Urban Art and then the work that they are doing now in university.

One student is studying neuroscience and the other civil engineering but technology is still a key part of what they are doing today.It was interesting hearing about their paths. One thing that became clear is that Urban Arts really promotes teamwork which is a critical soft skill. Students also talked about the encouragement they got from Urban Arts to look at other areas of student and broaden their experience. That is something that good teachers do!

Obviously, the keynote was a promotional talk for Urban Arts but it was also look into the evolution of a program and a good story of the value of cross curriculum education.

For my first session I decided to attend one on Figma. Figma is totally new to me so I was curious. Figma is a design tool that is apparently being used widely in industry. Figma has education programs (free) that they make available. The presenters were both former teachers which was encouraging to me.

This was basically a “sales pitch” for their education program and the resources that are part of it. Like a lot of companies, they would like to see students get to university or internship with a knowledge base in their products.That is not a bad thing and I don’t mean it as a criticism. I can see how using a design tool like this could be a valuable addition to a curriculum. Being free is of course a plus.

Next session was on teaching Ethics in AI based on the CSTA standards. (Slide deck at dayofai.org/cstany ) This presentation was by Day of AI dot org.

Day of AI has a lot of resources for teaching AI. They include a trivia game, individual lessons, and a full semester course. Materials are arraigned by grade level and range from Pre-K through high school. The youngest grades only have a few lessons and as the grade levels increase the number of lessons increases. There is a lot of focus on fairness and why AI may not always be fair.

I love that they emphasize the need for human judgment. They talk about lot about AI sometimes lacks context, full information, or understanding of right and wrong. These are materials well worth checking out.

After lunch was my presentation. I talked about Subgoal Labeling, Parsons problems, Live Coding, and a little on rubber duck debugging. My PowerPoint is at http://www.acthompson.net/TeachingCS.pptx and a PDF version is at http://www.acthompson.net/WhatWorks.pdf if you are interested. Maybe someone else will review it.

After my session, I went to a session about a post AP project based learning effort using AI to create a complete web application. Students created an online ordering app for a coffee shop. They used things from multiple CS courses and worked in a team. There was a solid planning phase, a research phase, a learning phase, and a build phase when students could use AI to help develop the site. It was a really good project.

Last session of the day,From Python Pixels to Prompts. The software used in this session is from Imagilabs.com software. This was a middle school targeted presentation but I really like the art aspect. The sire does have resources for grades 3 through high school. In this session, we were shown how students use Python code to draw images using a 8 by 8 grid of “pixels.”  Students learn the basics of programming and have fun. I like the creative aspect of using simple images as the output.

Part two of the talk was about having students prompt AI to write code using Vibe Coding ideas. They work a lot of planning into the process before the AI gets involved. I think that is an important idea. Planning makes using AI much more effective.

One statement that was made (I think during the keynote) was

Code is not the opposite of design. Code is the material for design .

That wasn’t the official theme of the conference but it was an idea that threaded its way though the sessions I attended. Overall, a very good conference. Lots of opportunity to connect with other educators and all of the sessions I attended were high quality.

Saturday, November 07, 2026

AP CS Principles Redesign–Good, Bad, or just Different?

The College Board recently announced a AP Computer Science Principles Redesign that has got a lot of people talking. I read through the preview framework and it actually looks pretty good to me. More on  that in a bit but first the things in the announcement that are causing the discussions.

The first big change is that there is heavy AI integration required in the course. In fact, the announcement says:

  • Students will complete a required series of hands-on, structured AI lab experiences.

and

  • Students need access to a district-approved AI tool that accepts a student’s own written instructions and returns a response. Schools will need to make sure every student has access to one.

In short, students must use AI and the school must make it available. I imagine the AI companies are pretty happy about that. One university professor made the comment on Facebook “APCSP sponsored by OpenAI. With chatAPCSP available at a discount”

Some school administrators, school IT people, and a lot of teachers in a range of domains may not be too happy about that. On the other hand, is it really possible to do a good job of computer science education without using some AI tools? Probably not.

Possibly the most controversial part of the redesign though is the AI Create Project. Students will be required to use AI for their create project. Is this because someone assumes that students will use AI anyway so let’s just make it legal? I doubt anyone would admit to that even if it is true. And maybe it’s not but it feels that way to me. My worry is that students will wind up believing that they don’t have to write any code at all and will let AI do it for them. I remain convinced that doing it the hard way is the best way to learn. We’ll see I guess.

There is an emphasis in the framework on students reading and evaluating the AI generated code. How well they will be able to do that without having experience writing code on their own (without AI help) is not something that I can predict.

Another important piece to the redesign is that starting in the 2028/2029 school year a text-base language will be required for the Create AI. I believe that the timing is to allow teachers more time to learn a text based language and develop teaching methods and resources to teach it. A lot of teachers are only using block based languages these days.

Some curriculum providers are already moving to text based languages. I understand that CodeAI (nee code.org) is moving the Python for example. Teachers should look into the plans of any curriculum and/or tools suppliers they are using to find out  their plans and timings.

Looking at the AP Computer Science Principles Course effective fall 2027 I do think it is basically sound. I was afraid that they were going to lose a lot to make room for AI. There is an optional and untested Unit 5 that has a lot of what I was afraid would be lost. I wouldn’t keep that optional but a lot of schools may have trouble fitting it all in. Especially districts with fewer than 180 days of instruction.

I will be at the CSTA New York Conference next week (14 November 2026) and if you are there please find me and tell me your thoughts on the redesign. Or just to say hi.

Thursday, November 05, 2026

Google AI, Product lock In, and Schools

Apparently, Google announced that schools would be able to block Google’s AI tools. The catch is that doing so will mean they cannot use Google Search. My first reaction, like many people was “so what?” There are plenty of other good search engines out there. I generally use Bing. I know people whose first choice is Duck Duck Go.

There are deeper concerns though.

Many schools are deeply tied in with the Google ecosystem. If they can use search to (attempt to) force schools to allow their AI what else can they use as leverage? One has to ask if dependence on one major tech company is a good thing for schools? I don’t just mean Google either. Microsoft and Apple are also building platforms for schools to use. None of this, not even so-called free tools, is completely altruistic. As students grow up and out of school they tend to take the tools they learned to use in school into their adult life.

Now I am not against technology in schools. I am also not against using software from large tech companies. What I am for is balance and diversity. I know that some people are going to use Google’s actions as a reason to move to open source. I’m not opposed to that either. Although, I have reservations about open source products as well. Different reservations than what I have about commercial products but different is not always better or worse. It can just be different.

Ultimately, schools and educators have to decide what is best for the students. What tools help students learn the most in the best ways possible. If a company tries to force (they might say “encourage”) schools to do things that school leaders are uncomfortable with doing than they have options. Now is as good a time as any for schools to start looking at them.