What Should Computer Science Undergraduates Study?

I noticed that Zhihu was pushing unrelated content to me, so I searched for computer‑related questions and found the query "What should be learned during a computer‑science undergraduate program?" and decided to write a broadly applicable answer for most graduates.

Having spoken with university teachers and students, I will share my perspective based on those conversations and my own experience.

Many universities teach material that deviates from industry needs; students waste time on subjects that are not required for employment, such as memorizing TCP header fields, Java 11 new features, or C++17 additions—information that can be easily looked up online.

Domestic universities often favor fixed, quantifiable exam content because it simplifies grading, allowing professors to relax or focus on other matters.

Some instructors are not dedicated to teaching; for example, a teacher who has taught Java for over a decade could not explain the usage of the transient keyword when asked.

Thus, a debate has emerged online about whether universities should expand enrollment.

While many argue that students should master abstract, core concepts like operating‑system principles and Linux kernel scheduling, I acknowledge their importance for building a solid foundation, yet for the average student they offer limited immediate benefit.

If you invest heavily in OS, networking, and algorithms while a roommate focuses on Java, multithreading, SpringBoot, and cloud frameworks, the roommate will likely progress faster, gain more practical experience, and have a richer résumé for job hunting.

Do not assume that learning low‑level fundamentals alone grants superiority; most development work consists of CRUD tasks, and the ability to deliver functional results is crucial, as companies hire to extract value at the lowest possible cost.

Without exposure to databases, you cannot grasp transaction mechanisms; without writing Java or C, you cannot understand object‑oriented or procedural paradigms. A basic familiarity is necessary before deeper concepts become clear.

Therefore, if you are not inclined toward deep theoretical study, a pragmatic, "ready‑to‑use" approach is acceptable—open‑source tools accelerate productivity, though building your own solutions can still be valuable when appropriate.

The reality is that efficiency and value creation are rewarded; while this utilitarian mindset may seem harsh, it aligns with market demands.

Abstract theory is beautiful, and engineering is beautiful; after reading extensive textbooks you may appreciate the clean logic of theory, yet it does not immediately generate economic value, leading many to prefer refactoring messy code only when performance matters.

In summary, what you should study in undergraduate years depends on your goals: if you aim for employment, focus on industry‑required skills, enrich your résumé, and master practical tools; if you cherish abstract concepts, consider graduate research.

I regret not providing many book recommendations or technical summaries, but I hope this guidance helps reduce confusion for fellow students.

Finally, I conclude with a meaningful closing remark.