Computing › Computer Science course guide
Computer Science: a guide for anyone thinking about taking the degree
What you study, where it came from, what helps to know beforehand, what changes with AI and where the degree leads. With exercises so you can test the ideas right here.
- 7 questions
- 13 interactions
- 8 myths
- 1 checklist to start today
The science of solving problems with steps a machine can follow
Computer Science is not the study of computers, in the same way astronomy is not the study of telescopes. The computer is the instrument. The object of study is the process: how to describe, step by step, the solution to a problem so that it always works, for any input, using as little time and memory as possible. That step-by-step has a name: algorithm.
Three questions run through the entire degree. What can be computed, and what cannot? How do you compute well, with the least possible effort? And how do you build real systems that do this for millions of people at the same time, without going down? The first belongs to theory; the second, to algorithms and data structures; the third, to systems, networks, software engineering, data and artificial intelligence.
I picked a number from 1 to 100. After each guess, I tell you whether the number is higher or lower. How many tries do you need?
Make your first guess.
Tries: 0Numbers still possible: 100
What is behind it. If you always guess the middle of the range that is left, you cut the problem in half at every step. With 100 numbers, 7 guesses are enough in the worst case; with 1 million, 20 are enough. Guessing in order (1, 2, 3…) can take 100 tries. Both approaches solve the problem; only one solves it well. This algorithm is called binary search, and it is this kind of reasoning, more than any programming language, that the degree teaches.
The areas of Computer Science
The degree covers all of them, and at the end you choose where to go deeper. Click an area to see the question it tries to answer and where it shows up in your life.
What can a computer solve, and what can it not?
It studies mathematical models of computation, such as the Turing machine, and proves limits: there are problems no program can solve, however fast the computer. It is the most mathematical part of the degree and the foundation for all the others.
In your life: it is thanks to this area that we know no antivirus can detect every possible malicious program.
What is the most efficient way to solve this problem?
It teaches how to organize data (lists, trees, graphs, hash tables) and how to measure how much time and memory a solution uses as the problem grows. It is the heart of the degree.
In your life: the route in your maps app is a shortest-path algorithm running on a graph of streets.
How does a person write instructions a machine understands?
It studies how programming languages are designed and how a compiler translates the text you write into the instructions the processor executes.
In your life: every app on your phone went through a compiler before it reached you.
What happens inside the machine when a program runs?
From the processor and memory to the operating system, which splits processor time among dozens of programs open at the same time.
In your life: it is the operating system that lets you listen to music while the browser loads a page.
How do many computers work together as if they were one?
It studies communication between machines and systems spread across many of them, which have to keep working even when part of the network fails.
In your life: your message arrives even if one of the app’s servers goes down along the way.
How do you store and find information without losing anything?
Data modeling, query languages and techniques so that millions of simultaneous operations do not corrupt the information.
In your life: your bank balance does not go wrong when two purchases happen in the same second.
How do you build large software, as a team, that works and can evolve?
Requirements, architecture, testing, quality and work processes. It is the bridge between the algorithm that works on paper and the system that works in production for years.
In your life: the app that gets an update every week without breaking what was already there.
How do you make a machine learn from data and make decisions?
Machine learning, neural networks, natural language processing and computer vision, with a lot of statistics and linear algebra underneath.
In your life: video recommendations, machine translation and generative AI assistants.
How do you protect systems and people from those who want to attack them?
Cryptography, authentication, vulnerability analysis and incident response. It requires thinking like the attacker in order to build the defense.
In your life: the padlock in your browser and two-step verification on your account.
How do you make technology people can and want to use?
It brings together computing, design and psychology to study how people interact with systems, including accessibility for people with disabilities.
In your life: the screen reader that lets a blind person use a phone.
How do you turn mathematics into images and motion?
Geometry, simulated physics and real-time rendering. It is where many people first reach computing, through games.
In your life: every frame of a game is computed dozens of times per second.
How do you use computing to do science in other fields?
Simulations and data analysis in biology, climate, physics, medicine and economics. Here the computer becomes a laboratory instrument.
In your life: the weather forecast is the result of numerical simulations of the atmosphere.
Computer Science and the other degrees in the field
In Brazil, the Ministry of Education’s curriculum guidelines for the field (CNE/CES Resolution No. 5, 2016) recognize five bachelor’s degrees and one teaching degree. The names look alike; the focus does not. [1]
- Focus
- The foundations: algorithms, theory, systems and the mathematical basis of computing.
- Question that drives the degree
- How and why does this work, and how can it be done better?
- Typical length
- 4 years
- A good fit for those who
- Want to understand computing as a whole and have the freedom to go into industry, research or their own business.
- Focus
- Hardware and software together: circuits, electronics, embedded systems and the computing that lives inside cars, machines and sensors.
- Question that drives the degree
- How do you design the machine and the program that runs on it?
- Typical length
- 5 years
- A good fit for those who
- Like physics and electronics as much as programming.
- Focus
- The process of building quality software at scale: requirements, architecture, testing, management and evolution of systems.
- Question that drives the degree
- How does a team deliver a reliable system, on time, and keep it alive for years?
- Typical length
- 4 years
- A good fit for those who
- Want to build products and enjoy organizing the work of many people.
- Focus
- Technology in the service of organizations: management systems, data for decision-making, business processes.
- Question that drives the degree
- How does technology solve the problem of this company or this industry?
- Typical length
- 4 years
- A good fit for those who
- Like technology and also management, business and people.
- Focus
- Training computing teachers for primary and secondary education, with the same technical base plus teacher training.
- Question that drives the degree
- How do you teach computing to children and teenagers?
- Typical length
- 4 years
- A good fit for those who
- Like computing and teaching.
- Focus
- Short-cycle technology degrees, such as Systems Analysis and Development or Computer Networks: shorter and more applied training.
- Question that drives the degree
- How do I make this work on the job, right now?
- Typical length
- 2 to 3 years
- A good fit for those who
- Need to enter the job market sooner and want practical training. You can deepen the foundations later.
Typical lengths. Each institution sets its own in the course’s pedagogical plan.
A science older than the computer
The idea of an algorithm is more than a thousand years old. In 1936, before any electronic computer existed, Alan Turing described an imaginary machine able to carry out any mechanical procedure and proved that there are problems no machine can solve. [2] Computing was born as a question in mathematics and logic; the engineering that built the machines came soon after. The first Computer Science department in the United States was created at Purdue in 1962, and Brazil’s first bachelor’s degree, at Unicamp, in 1969. [3,4]
Al-Khwarizmi and the word “algorithm”
The Persian mathematician wrote step-by-step procedures for calculations and equations. The Latinized form of his name gave rise to the word algorithm.
Ada Lovelace publishes the first program
In her notes on Charles Babbage’s Analytical Engine, which was never built, she described a procedure to compute Bernoulli numbers and realized that the machine could manipulate symbols, not just numbers.
George Boole turns logic into algebra
True and false become 1 and 0, and reasoning becomes arithmetic. Almost a century later, this algebra would become the language of digital circuits.
Alan Turing defines what it means to compute
In the paper “On Computable Numbers”, Turing describes an abstract machine that reads and writes symbols on a tape and shows that it can carry out any mechanical procedure. He also proves that some problems cannot be solved by any machine. It is the theoretical starting point of the field.
The program moves into memory
John von Neumann’s report on the EDVAC describes a computer that stores its program in the same memory as its data. Almost every computer today, including your phone, follows this idea.
The ENIAC and its programmers
One of the first general-purpose electronic computers is shown to the public. It was programmed by a team of six women mathematicians, who learned the machine from its diagrams, with no manual and no programming language.
Grace Hopper and the compiler
She creates the A-0, one of the first compilers: a program that translates instructions closer to human language into machine language. The idea paved the way for the languages we use today.
Artificial intelligence gets its name
A meeting of researchers at Dartmouth College establishes the term “artificial intelligence” and poses the question the field still pursues: can intelligence be described so precisely that a machine can reproduce it?
The first Computer Science department is born
Purdue University creates the first academic Computer Science department in the United States. Computing stops being a support service and becomes a discipline of its own. [4]
Brazil’s first bachelor’s degree in Computer Science
Unicamp creates the program, which graduates its first class in 1972 and becomes a model for many others in the country. The same year, in the United States, ARPANET connects the first computers of the network that would give rise to the internet. [3]
The Patinho Feio (“Ugly Duckling”)
At the Polytechnic School of the University of São Paulo, graduate students and professors build the first computer designed and manufactured in Brazil, as the final project of a Computer Architecture course. [5]
The Web opens to the world
Tim Berners-Lee, at CERN, makes the World Wide Web public: pages connected by links, accessed through a browser. The internet leaves the labs and enters people’s homes.
The Transformer architecture
The paper “Attention Is All You Need” introduces the neural network architecture behind today’s large language models.
Generative AI reaches the general public
Conversational assistants based on large language models start being used by millions of people to write, study and code. The question in section 04 of this guide starts here.
Everything becomes 0 and 1
The most basic foundation of the field is the bit: the smallest unit of information, on or off. Click the bits below and see the number and the letter they form.
What is behind it. Each bit is worth twice the one to its right. With 8 bits, one byte, you can represent 256 different values, from 0 to 255. A convention, such as the ASCII table, says that 67 is the letter C. Text, photos, music and video are, in the end, sequences of bits with an agreed way of reading them.
The pillars of the degree
Programming is the tool. What the degree builds is the ability to look at a problem nobody has solved yet and describe, precisely, how to solve it.
More curiosity than talent: what helps and what is not a requirement
Nobody needs to arrive knowing how to program. The degree starts from scratch, and many of the best students wrote their first line of code in the first week of class. What matters most is a solid high school foundation and the willingness to stick with a hard problem.
- High school math with no gaps: functions, exponents, logarithms, percentages and logical reasoning.
- Reading carefully: half of all programming mistakes come from a misread problem statement.
- English for reading: documentation, forums and papers in the field are, for the most part, in English.
- Patience with mistakes: programming is getting it wrong, understanding why and trying again.
- Studying on your own: the field changes fast, and much of what you will use ten years from now does not exist yet.
- Already knowing how to program. The degree teaches it from the start.
- Being a math “genius”. Enjoying it helps; you do not need to be top of the class.
- Having a powerful computer. An ordinary laptop handles most of the degree.
- Liking games. Many people arrive through games; many arrive by other paths.
- Having taken a technical course. It helps at first, but the difference disappears in one or two semesters.
Eight statements, no right answers. The test does not measure ability; it measures affinity with the kind of challenge the degree brings.
I like to understand how things work inside, not just use them.
When I get stuck on a problem, I keep at it until I find a way out, even if it takes time.
I enjoy puzzles, logic games, sudoku or chess.
I can explain a complicated idea in simple steps.
Math does not scare me, even when I do not get it right the first time.
I read tutorials or watch videos in English without much trouble.
I have built something on my own: a website, a mod, a clever spreadsheet, a robot.
I like working with other people to solve a problem.
Answer all 8 statements to see the result.
What the curriculum usually looks like, year by year
The base: learning to program and to think precisely.
- Introduction to Programming
- Discrete Mathematics
- Logic for Computing
- Calculus
- Linear Algebra and Analytic Geometry
- Introduction to Computing
- Object-Oriented Programming
The core: organizing data and understanding the machine.
- Data Structures
- Design and Analysis of Algorithms
- Computer Architecture
- Probability and Statistics
- Databases
- Programming Language Paradigms
Systems and theory: how everything runs and how far you can go.
- Operating Systems
- Computer Networks
- Theory of Computation
- Compilers
- Software Engineering
- Artificial Intelligence
Going deeper: you choose your path.
- Distributed Systems and Cloud Computing
- Information Security
- Machine Learning
- Electives
- Internship
- Final Project
Example of a typical bachelor’s curriculum, based on the Brazilian Computer Society’s training guidelines. Each institution designs its own. [6]
If AI already writes code, why study Computer Science?
Because writing code was never the hardest part. The hard part is knowing what to build, breaking the problem into pieces, choosing the right structure, predicting where the solution will fail and checking whether the result is correct. Generative AI sped up the first part a lot and left the others more exposed.
Stack Overflow annual developer survey, 2025. [7]
The numbers tell the same story you see in classrooms and companies: almost everyone uses it, and those who use it a lot have learned to be wary. The most common complaint is the almost-right answer, which seems to work and hides a subtle error. Those who spot the “almost” are those who know the fundamentals.
We asked an AI assistant for a function that computes a student’s average grade and says whether they passed (average of 7 or more). The code runs without any error message. Click the wrong line.
Which line would you change?
Why it matters. With grades 10, 6 and 6, the real average is 7.33 and the student passes. The function adds only 6 + 6 = 12, divides by 3, gets 4 and fails the student. The program does not crash, does not warn you and is silently wrong. This is the kind of error only someone who reads the code understanding each line will catch. And there is a second hidden problem: if the list is empty, line 5 divides by zero.
The bigger risk for beginners is not AI getting it wrong. It is AI getting it right in your place while you are still learning. The effort of solving a problem is what builds the judgment that, later on, lets you evaluate a machine’s answer. This argument is developed in the essay A camada que ensina (The layer that teaches: compression, abstraction and the formation of judgment in software engineering), published on the blog (in Portuguese).
- Try first, ask later. At least 20 minutes of your own attempt before asking for help.
- Ask for explanations, not just answers. “Why does this work?” teaches more than “solve it for me”.
- Always verify. Run it, test with hard cases, read it line by line. You are responsible for the result.
- Specifying well the problem before asking for the solution.
- Evaluating solutions: knowing whether an answer is correct, secure and efficient.
- Architecture: deciding how the parts of a system fit together.
- Understanding the domain: health, finance, education. AI does not know your client’s problem.
- Ethics and responsibility for what the system does to people.
What is not in the curriculum, but shapes the career
Two people with the same degree and the same grades often follow very different paths. The difference is almost always in skills that do not show up on a transcript. Click each card to see why it matters and how to practice it now.
Myth or fact?
Many people give up on the degree before starting because of a wrong idea. Answer each statement and see the explanation.
“To study Computer Science, you have to be a math genius.”
The math in the degree is demanding, but it can be learned through study, like any other subject. What it asks for is method and consistency. Enjoying it helps; genius is not a prerequisite.
“Computer Science is a programming course.”
Programming is one of the tools of the degree. It also covers theory, algorithms, mathematics, architecture, networks, data and engineering. If you only want to learn a language, there are shorter paths.
“AI will put an end to the profession.”
AI is changing the work: repetitive tasks get faster, and specifying, verifying and deciding gain weight. Someone has to build, evaluate and run these systems, and that takes exactly the training the degree provides.
“Computing is a man’s thing.”
Ada Lovelace wrote the first program, Grace Hopper created one of the first compilers and Margaret Hamilton led the software for the Apollo missions. And more women are joining every year: in Brazil, they already make up about a quarter of students entering computing degrees, compared with 15% of those graduating today. [8]
“People who study computing fix computers and printers.”
Equipment maintenance is a different kind of training. The degree prepares you to create systems, algorithms and solutions, not to repair hardware. But be ready: your family will ask anyway.
“It is a lonely field, for people who spend all day alone in front of a screen.”
Large systems are built by teams, with meetings, code reviews and conversations with users and clients. Knowing how to work with people is a central part of the craft.
“You can learn to program on your own, with short courses and practice.”
You can, and many people enter the job market that way. The degree adds what is hard to learn alone: the foundations of theory, algorithms and systems, which make a difference on the hardest problems, in research and in the ability to adapt when technology changes.
“With a powerful enough computer, every problem has a solution.”
Turing proved in 1936 that there are problems no program can solve, such as deciding, for any program, whether it will ever stop. And there are problems that have a solution but would take longer than the age of the universe to solve by brute force. Recognizing these cases is part of the training.
Where the degree leads
Demand for technology professionals in Brazil has been higher than the number of people graduating. According to Brasscom, the association of companies in the sector, the market needed about 665,000 professionals between 2019 and 2024, while about 465,000 graduated between 2018 and 2023. [9]
High demand does not guarantee a job for everyone: openings increasingly ask for experience and solid foundations, and competition for entry-level positions is real. But the degree opens up a rare range of paths. Click one of them.
Design, build and maintain systems
Apps, web platforms, systems for banks, hospitals and governments. From the first architecture sketch to fixing the problem that showed up at three in the morning.
Courses they use most: Data Structures, Software Engineering, Databases, Distributed Systems.
First step: build a complete project and publish the code.
Turn data into decisions and into models that learn
Build the pipelines that collect and organize data, analyze patterns, and train, evaluate and deploy machine learning models.
Courses they use most: Probability and Statistics, Linear Algebra, Databases, Machine Learning.
First step: take a public dataset and answer a question with it.
Protect systems, data and people
Find vulnerabilities before attackers do, design defenses, investigate incidents. This includes the ethical hacker, hired to test a company’s defenses.
Courses they use most: Networks, Operating Systems, Cryptography, Information Security.
First step: join beginner “capture the flag” security challenges.
Keep the systems the world uses up and running
Design cloud infrastructure, automate software delivery, plan capacity and make sure everything keeps working when demand multiplies.
Courses they use most: Operating Systems, Networks, Distributed Systems.
First step: put one of your projects online, reachable over the internet.
Create new knowledge and train people
Master’s and PhD programs, research labs at universities and companies, teaching. It is the path for those who want to work on questions that do not have answers yet.
Courses they use most: all of them, with emphasis on Theory, Algorithms and Scientific Methodology.
First step: look for undergraduate research early on. There is more about the research path on the Advising.
Create worlds, simulations and visual effects
Game engines, simulated physics, rendering, virtual and augmented reality. It mixes heavy math with creativity.
Courses they use most: Computer Graphics, Linear Algebra, Data Structures, Human-Computer Interaction.
First step: make a small game from start to finish and publish it.
Decide what is worth building
Product management sits between the user, the business and the engineering team. Those with a computing background speak the same language as the engineers.
Courses they use most: Software Engineering, Human-Computer Interaction, Databases.
First step: pick an app you use and write down what you would change in it, and why.
Turn a problem into a business
Computing is one of the cheapest fields in which to start a company: one person with a laptop can build the prototype. The hard part is everything else, and everything else can be learned.
Courses they use most: all the technical ones, plus what is not in the curriculum: selling, negotiating, leading.
First step: see the five steps just below.
From code to business, in five steps
Start with the problem, not the technology. Who suffers from it? How much does it cost them? Have they tried to solve it? A good business starts with a real problem, of real people, that you can describe in one sentence.
Build the minimum that solves it. A simple first version, sometimes without a single line of code, just to test whether the idea works. This is where technical training makes a difference: you build it yourself, and fast.
Put it in the hands of those who have the problem. Observe, measure, listen. Most ideas change at this stage, and that is good: it is the scientific method applied to a business.
Find out who pays and why. A useful solution nobody pays for does not last. Define the price, the cost of serving each customer and how customers will find the product.
Grow without breaking. More customers call for systems that handle the demand, a team, processes and, often, investment. This is where distributed systems and cloud computing stop being courses and become survival.
Innovation ecosystems help a lot along the way. Porto Digital, in Recife, is one example: it brings together companies, startups, universities and research centers in the same area. It was from there that Carlos Diego founded Valcann, a cloud computing company that, in 2023, became part of EPI-USE, of Group Elephant. The full story is in the Profile.
The questions we hear most
What is the difference between Computer Science and Information Systems?
Computer Science goes deeper into the foundations: algorithms, theory, systems and mathematics. Information Systems focuses on technology applied to organizations and business. Both train good professionals; the choice depends on whether you want to understand computing from the inside or apply it to management problems. The comparison in section 01 shows the differences side by side.
Do I need to know how to program before I start?
No. The degree teaches from the very first line. Some experience helps in the first weeks, but the difference fades quickly. If you want to arrive with some idea, the “Start today” section shows where to begin.
How long does the degree take?
The bachelor’s degree usually takes four years, in eight semesters. Computer Engineering usually takes five. Short-cycle technology degrees take two to three years.
Can I study online?
There are online programs. Since the training involves a lot of practice, lab work and teamwork, it is worth checking carefully how each program organizes these activities. In any format, check e-MEC, the Brazilian Ministry of Education’s official registry, to see whether the institution and the program are accredited.
What computer do I need?
An ordinary laptop handles most of the degree. Many heavier activities run in the institution’s labs or in the cloud. More important than the machine is having regular access to it to practice.
Will I get a job?
Demand for professionals in the field is above the number of graduates in Brazil, as the figures in section 07 show. But the market chooses people with solid foundations, projects and initiative. Internships, undergraduate research and personal projects during the degree make a big difference.
I am over 30. Is it too late?
No. Many people come to computing from other fields, and that prior experience often becomes an advantage: someone who knows an industry’s problem and knows computing can solve what few others can.
And after graduating: master’s and PhD?
They are the path to research and to teaching in higher education, and they also open doors in company research labs. In a master’s you learn how to do research; in a PhD, you produce new knowledge. The Advising page explains what that journey is like.
Eight things you can do before college
Check what you have done. Your progress is saved in this browser.
0 of 8 done
Ask. No question about the degree is a silly one.
Send your question through the contact form, or read the technical posts and the essays on education on the blog.
- Brazil. National Council of Education. CNE/CES Resolution No. 5, of November 16, 2016: National Curriculum Guidelines for undergraduate programs in Computing. abmes.org.br
- Turing, A. M. On Computable Numbers, with an Application to the Entscheidungsproblem. Proceedings of the London Mathematical Society, s2-42, 1936.
- Agência FAPESP. Unicamp’s Institute of Computing reaches 100 theses and 600 dissertations (on the creation, in 1969, of Brazil’s first bachelor’s degree in Computer Science). In Portuguese. agencia.fapesp.br
- Purdue University, Department of Computer Science. History of the Department (first Computer Science department in the US, October 1962). cs.purdue.edu
- Revista Pesquisa FAPESP. The legacy of the Patinho Feio (2022). In Portuguese. revistapesquisa.fapesp.br
- Brazilian Computer Society (SBC). Training Guidelines for Undergraduate Programs in Computing (2017). In Portuguese. books-sol.sbc.org.br
- Stack Overflow. 2025 Developer Survey, AI section. survey.stackoverflow.co
- Brazilian Computer Society (SBC). Paper in the Proceedings of Women in Information Technology (WIT 2024) with data from the 2022 Higher Education Census: women were 25.5% of entering students and 15.3% of graduates in computing programs. In Portuguese. sol.sbc.org.br
- Brasscom, via Convergência Digital. ICT may create up to 147,000 formal jobs in Brazil in 2025 (demand 2019–2024 and graduates 2018–2023). In Portuguese. convergenciadigital.com.br