Grade 10: Computer science

In this unit, 'Python: functions, files and data structures', you learn to write programs using lists, dictionaries and files. You work with notes, a worked line and a self-mark. You see it, say it, then do it alone.

• See it: watch or try the skill in 'Python: functions, files and data structures' on a teacher screen or a safe device.
• Name the parts: write the key words and what each part does.
• Use it: write programs using lists, dictionaries and files. Save your work in a named file or folder.
• Test and fix: run it, find one fault, correct it, and note what you changed.

Step by step.

1. Step 1, Write a function: A learner writes def average(nums): return sum(nums) / len(nums). Calling average([72, 91, 65, 88]) adds 316 and divides by 4, returning 79.0. A function packages one job so it can be reused without rewriting.
2. Step 2, Use a dictionary: He stores marks = {'Learner 1': 72, 'Learner 2': 91, 'Learner 3': 65, 'Learner 4': 88}. marks['Learner 2'] gives 91. A dictionary looks up a value by its key, which is faster to read than searching a list by position.
3. Step 3, Write to a file: with open('scores.txt', 'w') as f: for n, m in marks.items(): f.write(n + ',' + str(m) + '\n'). The file has 4 lines such as Learner 1,72. The with block closes the file safely at the end.
4. Step 4, Read it back: with open('scores.txt') as f: lines = f.read().splitlines() gives a list of 4 strings. For each line, n, m = line.split(',') separates the name and text number, and data[n] = int(m) rebuilds a dictionary.
5. Step 5, Use a list and dictionary together: best = max(data, key=data.get) returns 'Learner 2', because Learner 2 has the highest value, 91. print(best, data[best]) displays Learner 2 91. print(average(list(data.values()))) shows 79.0, the same as before.
6. Step 6, Test an edge case: If scores.txt is missing, open raises FileNotFoundError. He wraps the reading in try and except FileNotFoundError and prints 'No file yet'. Handling errors stops the program from crashing.
7. Your turn: stock = {'apple': 5, 'pear': 3, 'plum': 4}. Write code for the total and the most stocked item. Model answer: total = sum(stock.values()) gives 12; best = max(stock, key=stock.get) gives 'apple'.

Activity. About 30 to 40 minutes of class work, then one short practice task for home. Keep evidence: a worked task and a corrected copy.

Standard to meet. Write programs using lists, dictionaries and files: at least 8 of 10 marks on a short task, with one line of reasoning for each answer.

In this unit, 'Databases and SQL', you learn to design and query a small database. You work with notes, a worked line and a self-mark. You see it, say it, then do it alone.

• See it: watch or try the skill in 'Databases and SQL' on a teacher screen or a safe device.
• Name the parts: write the key words and what each part does.
• Use it: design and query a small database. Save your work in a named file or folder.
• Test and fix: run it, find one fault, correct it, and note what you changed.

Step by step.

1. Step 1, Design two tables: A learner separates people from marks. students(student_id, name) and marks(student_id, subject, mark). student_id is the primary key in students and a foreign key in marks, so each mark points to one student.
2. Step 2, Create and fill: CREATE TABLE students (student_id INTEGER PRIMARY KEY, name TEXT); and CREATE TABLE marks (student_id INTEGER, subject TEXT, mark INTEGER); then INSERT students (1,'Learner 1'),(2,'Learner 2') and marks (1,'Maths',72),(1,'Science',80),(2,'Maths',91),(2,'Science',85). That is 2 students and 4 marks.
3. Step 3, Why two tables: Names are stored once, not repeated in every mark row. If Learner 2 changed their name, only 1 row in students changes, not 2 rows in marks. This avoids repeated data and mistakes.
4. Step 4, Join: SELECT s.name, m.subject, m.mark FROM students s JOIN marks m ON s.student_id = m.student_id WHERE m.mark >= 80 ORDER BY m.mark DESC; returns Learner 2 Maths 91, Learner 2 Science 85, Learner 1 Science 80. Learner 1's 72 fails the WHERE test.
5. Step 5, Group: SELECT s.name, AVG(m.mark) FROM students s JOIN marks m ON s.student_id = m.student_id GROUP BY s.name; returns Learner 1 76.0 and Learner 2 88.0. Check: Learner 1 is (72 plus 80) divided by 2 is 76; Learner 2 is (91 plus 85) divided by 2 is 88.
6. Step 6, Check for mistakes: A mark with student_id 3 would match no student, so the join would not show it. The learner runs SELECT COUNT(*) FROM marks; and gets 4, then checks the join returns the 4 matching rows.
7. Your turn: Tables books(book_id, title) and loans(book_id, days) with books (1,'Sea'),(2,'Sky') and loans (1,7),(1,5),(2,10). Average days per title? Model answer: join on book_id, GROUP BY title; Sea is (7 plus 5) divided by 2, which is 6.0; Sky is 10.0.

Activity. About 30 to 40 minutes of class work, then one short practice task for home. Keep evidence: a worked task and a corrected copy.

Standard to meet. Design and query a small database: at least 8 of 10 marks on a short task, with one line of reasoning for each answer.

In this unit, 'Artificial intelligence project cycle', you learn to plan a small data project from problem to evaluation. You work with notes, a worked line and a self-mark. You see it, say it, then do it alone.

• See it: watch or try the skill in 'Artificial intelligence project cycle' on a teacher screen or a safe device.
• Name the parts: write the key words and what each part does.
• Use it: plan a small data project from problem to evaluation. Save your work in a named file or folder.
• Test and fix: run it, find one fault, correct it, and note what you changed.

Step by step.

1. Step 1, Scope the problem: A learner's team asks, 'Which students are likely to return library books late?' The goal is a yes or no prediction so a librarian can send a reminder. They write the problem, who it helps and what a good result means.
2. Step 2, Acquire data: They use 250 anonymous loan records with 3 columns: number of books out, days since last visit, and returned late yes or no. Names are removed to protect privacy, and the data is used only for this project.
3. Step 3, Explore and prepare: They find 12 rows with empty values and remove them, leaving 238. They check balance: 175 returned on time and 63 late. If the model always guessed on time it would be right 175 of 238, about 73.5 percent.
4. Step 4, Split: They keep 188 rows for training and 50 for testing, since 188 plus 50 is 238. The model must be tested on rows it has never seen, otherwise it may only memorise the data.
5. Step 5, Evaluate: On the 50 test cases the model gets TP 18, TN 22, FP 4, FN 6. These total 18 plus 22 plus 4 plus 6, which is 50. Accuracy is (18 plus 22) divided by 50, or 80 percent.
6. Step 6, Precision and recall: Precision is 18 divided by (18 plus 4), which is 81.8 percent: of the students predicted late, most were. Recall is 18 divided by (18 plus 6), which is 75 percent: it caught three quarters of the late returns.
7. Your turn: A test has TP 10, TN 30, FP 5, FN 5. Find accuracy and recall. Model answer: total 50; accuracy is 40 divided by 50, or 80 percent; recall is 10 divided by 15, which is 66.7 percent.

Activity. About 30 to 40 minutes of class work, then one short practice task for home. Keep evidence: a worked task and a corrected copy.

Standard to meet. Plan a small data project from problem to evaluation: at least 8 of 10 marks on a short task, with one line of reasoning for each answer.

In this unit, 'Cyber law and ethics', you learn to describe privacy, copyright and responsible AI use. You work with notes, a worked line and a self-mark. You see it, say it, then do it alone.

• See it: watch or try the skill in 'Cyber law and ethics' on a teacher screen or a safe device.
• Name the parts: write the key words and what each part does.
• Use it: describe privacy, copyright and responsible AI use. Save your work in a named file or folder.
• Test and fix: run it, find one fault, correct it, and note what you changed.

Step by step.

1. Step 1, Privacy case: A learner's app asks for the user's location, contacts and photos just to show a school timetable. The learner asks, 'What data is really needed?' Only the timetable is needed, so asking for 3 data types is too much.
2. Step 2, Apply data minimisation: She keeps only what the function needs and removes location, contacts and photos. Collecting less data lowers the risk if the data leaks, and users must be told clearly what is collected and why.
3. Step 3, Copyright case: For a poster she finds a photograph online. The page says 'All rights reserved'. That means she may not copy it without permission. She can ask the owner, use a picture marked free to reuse with credit, or draw her own.
4. Step 4, Give credit properly: She picks an image whose licence allows reuse with credit, and writes under it: creator name, title, source website, and licence name. Credit does not replace permission when the work is not licensed.
5. Step 5, Responsible AI case: For a history essay she uses an AI tool to draft a paragraph. She checks each fact, rewrites in her own words, and tells her teacher she used AI for a first draft. She does not hand in the draft as her own work.
6. Step 6, Cyber law idea: Breaking into someone's account, spreading harmful content or sharing a classmate's private data can be offences under cybercrime and data protection rules. The learner checks the country's current rules or asks the teacher.
7. Your turn: Your team wants to use a song on a school video. What do you check? Model answer: check the licence or get permission, use free-to-reuse music with credit, or compose your own; do not just copy it.

Activity. About 30 to 40 minutes of class work, then one short practice task for home. Keep evidence: a worked task and a corrected copy.

Standard to meet. Describe privacy, copyright and responsible AI use: at least 8 of 10 marks on a short task, with one line of reasoning for each answer.

On the slide: L1 Basic · Computer science as problem solving

Notes: Name what makes a problem suitable for a computer, such as clear steps and repeated work. Notes: ask learners how they would find the tallest person in a room; they describe each step in words. No devices are needed, just a clear list.

On the slide: L2 Awareness · Algorithms and flowcharts

Notes: Describe an algorithm as a flowchart and trace it by hand. Notes: draw a flowchart that finds the larger of two numbers, then run it with 7 and 4 and with 5 and 5. Learners say what the chart does when numbers are equal. Work on paper first.

On the slide: L3 Foundation · Functions and control flow

Notes: Write a function with conditions and loops and test it with chosen inputs. Notes: write a factorial function; 5 factorial must give 120, and 0 factorial must give 1. Learners test one normal case, one edge case and one wrong input. Take screen breaks.

On the slide: L4 Practitioner · Data structures and binary search

Notes: Apply stacks, queues and sorted arrays to solve a task and predict the cost. Notes: a sorted list of 1,000 items needs at most 10 binary-search steps, since 2 to the 10 is 1,024. Learners trace a search for 37 in a list of 16 numbers and count the steps.

This course: Grade 10 Computer science is at this level. Its units include Python: functions, files and data structures; Databases and SQL; Artificial intelligence project cycle; Cyber law and ethics.

On the slide: L5 Specialist · Recursion and analysis

Notes: Analyse a recursive function, count its calls and improve it with memoisation. Notes: a naive Fibonacci of 5 makes 15 calls. Learners draw the call tree, spot repeated work, add a cache and show the calls falling to 9. Test with small inputs before large ones.

On the slide: L6 Advanced · Complexity and proofs

Notes: Evaluate algorithms by growth rate and prove a simple correctness claim. Notes: for n of 1,000,000, n squared is 10 to the 12, but n log2 n is about 2 times 10 to the 7. Learners prove a loop invariant for insertion sort and state the best and worst cases.

On the slide: L7 Super expert · Lower bounds and research design

Notes: Evaluate limits of computation and design a rigorous experiment on algorithms. Notes: any comparison sort of 10 items needs at least the ceiling of log2 of 10 factorial, which is 22 comparisons. Learners plan benchmarks with controlled inputs, repeats, confidence intervals and a data-sharing statement.