Week 2

Question 1

When looking at integration testing, what do we look at?

Answer 1

Incremental and non-incremental.

Question 2

When looking at performance testing, what do we look at.

Answer 2

Scalability, stability, speed and memory.

Question 3

What do we look for in algorithms?

Answer 3

We seek algorithms that are correct (ideally a proof) , efficient, and easy to implement.

Question 4

Will iterative cases always be faster than recursive, or the other way around?

Answer 4

No, it is not always the case that one will be faster than the other purely based on whether they are iterative or recursive.

Question 5

How do we determine if an application is too slow?

Answer 5

The application is too slow if it doesn’t meet your project’s stated performance requirements.

Question 6

What are some internal factors that could affect an algorithm performance analysis?

Answer 6

Runtime and memory requirement.

Question 7

What are some external factors which affect an algorithm performance analysis?

Answer 7

Input size, computer specs, compiler quality.

Question 8

How is complexity analysis computed?

There are three.

Answer 8

• Time complexity (The time required)
• Computational complexity (The number of steps or arithmetic operations)
• Space complexity (The amount of memory required)

Question 9

What are the issues with python’s time module? Why is timeit better?

Answer 9

The time module records the start and end. However, it can be effected by alot of outside factors and is considered quite noisy.

The timeit module is better since it helps eliminate non-functional, non-essential time consuming tasks. It is generally more accurate.

Question 10

What is profiling.

Answer 10

Profiling is measuring relative system statistics.

It tells us, where most of the time is being spent. (classical profiling)
Which method takes the most time, and which is called the most.

Profiling is not the same as benchmarking or optimizing.

Question 11

What is optimization? Why do we warn not to optimize too quickly?

Answer 11

Optimizing is refactoring and enhancing to speed up the code. Optimizing too quickly can lead to code that is barely used being unnecessarily optimized.

In general only optimize if you have a clear idea of what is actually slowing down your code.

Question 12

What is the Pareto Principle?

Answer 12

80% of a program’s execution occurs within 20% of its code.

Question 13

What is Big-O Notation? What is it used to specify?

Answer 13

Big-O is used to specify an upper bound on a function.

For a bound, choose the smallest simplest function that satisfies the inequality. Choose the inequality.

Question 14

True or false, Big-O ignores the constant in analysis, so g(n) is identical to g(n) = 2n.

Answer 14

Yes, Big-O ignores the constant. g(n) and g(n) = 2n are identical.

Question 15

What are the types of Algorithms?

Answer 15

Constant
Logarithmic
Linear
N-Log-N
Quadratic
Cubic
Exponential

Question 16

What is considered an efficient algorithm?

Answer 16

Quadratic algorithms an above are only practical with small inputs.

Anything log to linear are considered good operation times.

Question 17

What is Big-Omega used to specify?

Answer 17

It is used to specify a lower bound on a function.

Question 18

How should we choose the bound for Big Omega?

Answer 18

We should choose the largest simple function that satisfies the inequality.

Choose the tightest bound possible.

Question 19

What is Big-Theta used to specify?

Answer 19

It is used to specify both upper and lower bounds on a function.

Question 20

When is the only time that we can use Big Theta?

Answer 20

We can only use it if big-O and big-Omega have the same complexity.

Question 21

Which case do we usually care the most about? Best, worst or average case?

Answer 21

We usually care about the worst case. That is unless the worst case is very unlikely.

Then we would choose the average case.

Question 22

What do we ignore in complexity analysis?

Answer 22

Constants, addition, subtraction