Even after a year or more of programming instruction, students have grcat difficulty predicting what output a program will have, in what ordcr command will hc executed, or in writing and debugging original programs to solve problems.

Programming and coding are skill sets. Programmers can progress from beginners to experts over a long period of time. It takes more and more practice and experience to reach the higher stages. Becoming proficient at programming real-world applications is not easy. The much publicized kid coders are mostly beginners.

programming is a hard skill to learn. The lack of student programming skill even after a year of undergraduate studies in computer science was noted and measured in the early 80's [32] and again in this decade [20]. We know that students have problems with looping constructs [31], conditionals [8], and assembling programs out of base components [33]

Educators have been searching for many years for ways to accelerate learning programming. Seymour Papert introduced the Logo language in the 1970s and watched how children got excited by computing and learned how to think computationally. He attuned Logo to children’s interests; even so, it still took students time to move from the fascination of the introduction to the ability to program useful computations regularly.

Most of our studies point more toward how complex it is for humans to learn how to program a computer. We continue to marvel at the intelligence and creativity of humans, and that even students without training in computing can already think in algorithmic terms. However, developing that skill to the point that it can be used to give directions to a machine in the machine’s language occurs more slowly than we might expect. We can get students through the process, but we still don’t have effective measures of how much they’re learning.

The reality is that while writing code is hard, the harder part for students (and really anyone) is understanding how it executes and then making decisions about what it should do differently. Program comprehension is what makes APIs hard to use (because they intentionally hide what they do, capturing behavior only through poorly written natural language). It’s what makes programming languages hard to use (because debugging tools are so poor slowing down enough to teach). It’s what makes large software systems hard to change and evolve (because of the sheer amount of code to understand). LLMs do nothing to make this comprehension or decision making easier.