Week 3: How you measure an algorithm

Week 3 · 18 September 2026 · Data Structures (数据结构), Tsinghua University · XuetangX

All theory this week. Measuring algorithms without running them.

1. What I worked on

绪论 B1: performance measurement, problem size, worst-case analysis, and the idealised computation model.

2. What was challenging, and how I handled it

Writing cost as T(n) instead of a single number was new. For sorting, n is the array length; for multiplication, n is the digit count. Same letter, totally different meaning. I paused the video and worked through my own examples until it felt natural. Worst-case analysis I already knew from competitive programming (you always check the worst input), but the course treats it as a modelling choice, not just common sense. That reframe was useful.

3. How well I ran my own learning

Self-score: 100

I kept connecting the theory to stuff I already knew from competitive programming, which helped the abstract parts make sense. I paused videos to think through things on my own.

4. What I’ll do differently next session

After each video, close the tab and restate the main idea in one sentence before moving on.

Evidence

Notes on T(n), worst-case vs average-case, and the idealised model.