England: Time to replace Computer Science with Computing

This is policy wonk stuff, but I find policy fascinating. As a researcher, it’s hard to figure out “How are most people (students, faculty, whatever) in this field thinking about X?” Policy-makers have to figure that out, too, and then have to respond. A change in policy is like a research paper that says, “We found that the status quo wasn’t working anymore.”

The English government has just conducted an independent review of all their school curricula (see report here).  The review is critical of how Computer Science is working in English schools today. They say that Computing now pervades all disciplines and “digital literacy” should be taught in an integrated manner. I recommend reading the report — it’s accessible and covers a bunch of important issues, like who is taking CS and where there’s a split between policy and practice.

One of the explicit recommendations is that the government:

Replaces GCSE Computer Science with a Computing GCSE which reflects the full breadth of the Computing curriculum and supports students to develop the digital skills they need.

The government response (linked here) agrees:

We agree with the Review that the computing curriculum should be the main vehicle for teaching about digital literacy, and we are confident that delivering the computing recommendations will provide more pupils with valuable digital skills that are essential for the future.

It is also clear that, in some subjects, digital methods now influence the content and how it is taught. We will work with experts to assess the validity of digital practice in these subjects, the evidence of whether this can be done robustly and whether it merits inclusion in the new curriculum. Where it does, we will include a requirement for the relevant digital content in those subjects’ programmes of study and we will ensure that it aligns with the computing curriculum, to reduce the risk of duplication.

We will also replace the computer science GCSE with a broader offer that reflects the entirety of the computing curriculum whilst continuing to uphold the core principles of computer science such as programming and algorithms, and explore the development of a level 3 qualification in data science and AI.

Bottomline: CS just isn’t the thing anymore. Computing and computing across the curriculum is what is needed.

As a director of a Program in Computing for the Arts and Sciences, and someone who spent 25 years in a College of Computing, I wholly endorse this change and welcome it. As I described in a blog post from a couple of years back, “computer science” was originally invented to be a broad subject to be taught to everyone. Over the last 60 years, “computer science” has become more narrow (e.g., overly emphasizing algorithms while de-emphasizing building and creativity and social impacts, as Sue Sentance describes in this blog post, while “computing” represents a broader perspective. When we think about what should be taught to everyone in secondary school, Computing (and digital literacy, as the reports suggest) are more appropriate than what we now mean when we say Computer Science.

GenAI as automobile for the mind, and exercise as the antidote: A metaphor for predicting GenAI’s impact

Some of you may remember the Apple ads that emphasized the computer as a “bicycle for the mind.”

(From https://folklore.org/Bicycle.html)

GenAI is not like a bicycle for the mind. Instead, it’s more like an automobile. I’m finding that comparison to be useful in thinking about how GenAI may impact our world.

A bicycle extends our abilities. It allows us to do more with our legs and bodies than we can without the bicycle. The automobile also extends our abilities, but it doesn’t use those abilities. As Paul Kirschner recently wrote, GenAI is not cognitive offloading. It’s outsourcing. We don’t think about how to do the tasks that we ask GenAI to do. As the recent Anthropic study showed, you don’t learn about the libraries that your code uses when GenAI is generating the code for you (press release version, full ArXiv paper).

Automobiles have had an enormous impact on modern society. We can go places and do things that we couldn’t previously. Most of us can’t bike across the US, but many of us drive across it. So, we drive a lot and bicycle less. But there’s a cost — our bodies and minds atrophy if we do not use them.

Ford sold the Model-T as a general tool. (“You can have any color you want as long as it’s black.”). Users made changes to it to adapt it for various conditions and tasks. Today, those changes have led to a wide range of vehicles for a wide range of purposes: sedans, minivans, pick-up trucks, SUVs, and big rigs for hauling materials over long distances. Ford could not have anticipated all those different uses and specializations. They evolved over time.

To gain the benefit of automobiles, we made enormous changes to our infrastructure. We have freeways and driveways, garages and parking structures, gas stations and tire shops. We have changed our environment in order to use the automobile more.

But over time, we have seen those costs. Automobiles (and associated infrastructure, like blacktop parking lots) have had a large, negative impact on our ecology. Neighborhoods were destroyed when freeways were built through them.

We are starting to roll back some of our society’s earlier decisions that favored the automobile. We develop hybrid and electric cars that have less negative impacts on our ecology. In my town, bike lanes are being added, explicitly to choke automobile traffic in order to encourage more biking and less driving. We want people to use their bodies more. Exercise is an antidote to many of the automobile’s ills.

Here’s what I predict based on this comparison:

• We are going to see more specialized forms of GenAI, that we are going to have difficulty imagining today. Already, I am seeing the best learning outcomes from tools where GenAI is built into the tool (like Xu Wang’s work and Barbara Ericson’s). Chat is the Model-T of GenAI. It’ll get used for lots of purposes, but we’ll eventually figure out the specialized forms that will be much more useful.
• We are going to change our infrastructure to enable GenAI. More power plants, more power distribution, more data centers. Eventually, we’ll figure out that we went too far, and we’ll scale those back. But right now, it’s hard to estimate what’s “too far.”
• We will likely overuse GenAI and some of our abilities will atrophy, without a lot of self-regulation and careful consideration of how we use GenAI. Generative AI is a marshmallow test. We will have to figure out that we need to exercise our minds, even if GenAI could do it easier, faster, and in some cases, better.

Personally Meaningful Data to Motivate Learning in Data Science and AI

I have written several blog posts about the different ways to implement Media Computation in introductory programming courses. We built JES at Georgia Tech in 2002 and the final release was in 2020. Our introductory course in PCAS that uses Media Computation, COMPFOR 121: Computing for Creative Expression, uses Snap! and Pixel Equations (as described in this blog post). Our Python course (COMPFOR 221: Digital Media with Python) started in Python3 with the JES4Py library, but then we moved to Google Collaboratory notebooks (the libraries for that are available here).

Dave Largent at Ball State continues to teach Media Computation. The students in his course compete in an art show each term for which I’ve served as a judge. Dave let me know that he and his students have extended JES4Py and have released a new library:

I’ve had a couple of undergrad students working with me to redevelop/extend Gordon College’s JES4py package. We’ve published it at PyPI under the name mediaComp (https://pypi.org/project/mediaComp/). Our GitHub is https://github.com/dllargent/mediaComp.

Why is this interesting? Why is anybody teaching with a 20 year old method, and even making new libraries for it?

Maybe because it answers a CS education need that has only grown more important. Data science is a bigger deal now than it was 20 years ago. Ben Shapiro told us in 2018 that machine learning was going to change the CS curriculum, and we needed to think more about data. But what data are interesting to students?

The empirical studies in computing education research are pretty clear: Motivation matters.. Students get frightened off by computer science classes. They find our examples boring. If we can teach the same computing concepts using any data, why not use data that students find interesting?

I’ve heard Jens Mönig, lead architect and developer on Snap!, answer this question several times in several talks. There’s a new interview with him in the most recent ACM Inroads magazine (link) with Jens where he makes the point again. Students are interested in their data. Personal data, data about them, data that they make, data that are relevant to them. The phrase in the Constructionist community is “personally meaningful.”

Media Computation is data manipulation with personally meaningful data — your pictures and sounds, or the pictures and sounds that interest you. There are a lot of pixels and samples in those pictures and sounds. Those are data that matter to the students who care about those pictures and sounds.

Media Computation as an approach is not going to be for everyone. But every computing teacher should answer the meta question, “Why should my students care about these data?”. We often use the Corgis project data to help students find data that are personally meaningful. That’s where you can find the Titanic passenger dataset that Jens talks about in his interview. I am an advisor to API Can Code, which is a curriculum all about doing data science with live data that students might care about.

My point here isn’t that all teachers should use Media Computation. My point is that all computing teachers should engage students with personally meaningful data.

Student collage from COMPFOR 221 last year

Come join us at the ITiCSE 2026 Doctoral Consortium!

I’ve reached the stage in my career where I’m attending conferences not because I’m presenting a paper but because I’ve agreed to take on a service role.. That’s not a bad progression. I won’t be at ACM SIGCSE 2026 because I have neither a paper not a service role — I’ll miss seeing everyone, and hope you enjoy St. Louis.

I will be at ACM ITiCSE 2026 and 2027, serving as co-chair for the Doctoral Consortium. Please, PhD students, come join Monica Divitini from NTNU and me at this year’s DC in Madrid. More information on the DC and how to apply is here.

I will also be at ACM ICER 2026 and 2027, serving as co-chair for lightning talks and posters. I’ll pester you with that Call for Participation when the page gets posted.

Defining Learner-Centered Design of Computing Education: What I did on my sabbatical

My planned activity during my sabbatical was to revise my 2015 book “Learner-Centered Design of Computing Education.” One of the fixes I wanted to make was a better definition of what “learner-centered design” was. In the new edition, I wrote some formal defining stuff, and then I wrote the below — an extended metaphor to make distinctions between different kinds of “centering” in education. I’m sharing that section here (in its pre-reviewed and pre-edited state). It comes right after defining what the Zone of Proximal Development is and what student performance means.

There are many different kinds of teaching activity that can help a student reach a more sophisticated level of performance. A teacher can model successful performance. The teacher can give feedback on the student’s performance. The teacher can coach or guide a student while attempting a task. They can set expectations in the class which create a social context for success. They can use teaching methods that have a proven research record in promoting engagement and student performance.

Figure 1: A metaphor for teaching contrasting learner-centered and standards-centered

Consider teaching from the top or bottom of the ZPD. Here is a metaphor to make distinctions between two kinds of support in order to create a geography of teaching. Imagine the ZPD as a climbing wall (Figure 1). The student is at the bottom and wants to reach the top. Depicted as grayscale images in this figure, here are two ways a teacher might support the student in scaling this wall:

• The supporter at the bottom can help the student get started, giving them a “boost” or “leg up.”
• The supporter at the top can reach down, and get them the rest of the way to the top of the wall.

The supporter at the bottom is more flexible than the one at the top. She can move to where the student is actually standing. She can help the student scale different parts of the wall or even reach different goals along the wall. She can bend even further if the student is shorter.

But a disadvantage for the supporter at the bottom is that she cannot be absolutely sure that the learner reaches the top. She can meet the student where they are when they first face the wall. She can help them get started on whatever path they choose on the wall.

The supporter at the top can help students who are almost at the top of the wall. He can be sure that students actually reach the learning objective. When he is reaching down, he is in a fixed position. He can help the student reach the objective where he is at, the level that he has already achieved. He can also be sure when a student does not reach this standard – he can see the students who fall, or who do not make it to his level. He is in a better position to decide whether the student is going to achieve the desired objectives.

The supporter at the bottom is more learner-centered. The supporter at the top is more standards-centered. Neither supporter is particularly strong at helping the student in the middle, when the student is challenged to persist, to stay engaged, and to maintain motivation. If the student is not particularly interested in achieving the top of the wall, they are satisfied making it part-way to the objective, then the learner-centered teacher has the most to offer.

Learner-centered teaching is concerned with helping students where they are, helping them to get started, and getting them engaged and motivated to tackle the mid-part. Low enrollment and high withdrawal or failure rates (sometimes called WDF rates) are issues that learner-centered teaching addresses. Learner-centered teaching also addresses issues of diversity, with the goal that all kinds of students can succeed in the class — even those who think that they cannot succeed or do not have the prior background to succeed.

Standards-centered teaching is concerned about making sure that students have what they need to go on, in their studies or in their career. Students who fail the second class because they did not learn enough in the first class is an issue for standards-centered teaching. Talking to industry partners about the desired out- comes is standards-centered. Concern about what graduates can do and achieve is a standards-centered teaching issue.

(I’m skipping some text here about teacher-centered, classroom-centered, and other forms of structuring education.)

I am splitting hairs a bit between child-centered and learner-centered. Learner-centered also starts from the students’ interests and considers the learner’s needs, and is very much about student construction of knowledge in their own minds, since that is how learning takes place. As described in Chapter 2, the knowledge to be learned in learner-centered education is defined by the community of practice. That is external to the learner.

Within the metaphor, I am describing three kinds of teaching: Learner-centered (supporter at the bottom), standards-centered (supporter at the top), and maintaining motivation and engagement (in the middle). Of course, teachers and students have to address all these issues, but it is sometimes useful to focus on one part. Consider this metaphor: If you have heart problems, it is important to go to a cardiovascular specialist. That does not mean that you do not need to care about skeleton, digestion, and skin; you need all of those, but sometimes you can address critical issues or fix problems by specializing. I focus on the first one because it is the most important. I like the way my colleagues Amy Bruckman and Betsy diSalvo put it

Computer science is not that difficult, but wanting to learn it is.