Introductory Programming With Python

Introductory Programming with

PythonBrendan McCane

Hitchhiker’s Guide to Programming

Programming is hard. You just won’t believe how vastly, hugely, mind-bogglingly hard it is. I mean, you may think making a decent sandwich is hard, but that’s just peanuts to programming.

Donald Knuth says:

In fact, my main conclusion after spending ten years of my life working on the TEX project is that software is hard.

CS1

• Average failure rate for CS1 courses is 33% (min 0%, max 60%).

• Most universities teach Java, C# or C++.

• Industrially relevant and OO is more “natural” right?

Design Principles

• As simple as possible

• Few magical incantations

• Immediate feedback (no compile/run cycle)

• A practical and modern language

Target Audience

• Students majoring in computer science– Python in semester 1 (optional)– Java in semester 2 (compulsory)

• Students not majoring in computer science– Mostly science or technical majors

Curriculum

• We modified an open source textbook

• How to Think Like a Computer Scientist

• You can download our version from:– http://www.cs.otago.ac.nz/staffpriv/mccane/te

aching.html

Curriculum

• Introduction• Variables, expressions• Python builtins• Functions (2)• Conditionals• More functions• Test Driven Development• Files and modules• Iteration (2)

• GUI programming• Case study (2)• Strings (2)• Lists (2)• Tuples and sets• Dictionaries• System Programming• OOP• Case study

Course Structure

• Tight integration between lectures and laboratories

• 24 lectures, 21 laboratories

• Terms requirement – students had to submit 18 labs (not marked)

• Labs had basic + advanced exercises

• Lectures were programming demonstrations

Assessment

• Mid-semester test (20%) – multi-choice

• Final Exam (60%) – multi-choice

• Mastery tests (2 x 10%)– Programming problems under test conditions– Tests published beforehand– 4 sections per test, each worth 2.5%– Each section pass/fail– Graded based on doctests + code inspection

Mastery Test Example (lab 9)

def score(numbers):

"""

give the average of the numbers excluding the biggest

and smallest one

>>> score([2, 7, 9, 10, 13, 1, 5, 12])

7.5

>>> score([3, 7, 2.5, -4])

2.75

"""

Final Exam Example

def matrix_to_sparse(in_matrix):

sparse = {}

for row_index,row in enumerate(in_matrix):

for col_index,val in enumerate(row):

if val != 0:

sparse[(row_index,col_index)] = val

return sparse

matrix = [[0,0,1], [0,2,0], [3,0,0]]

sparse = matrix_to_sparse(matrix)

print sparse[(2,0)], sparse[(1,1)], sparse[(0,2)]

Outcomes

Outcomes

• 172 students attended first lab

• 28% A

• 23% B

• 22% C

• 27% failed (19% failed terms)

Notable Comments

• Easy topics: first half

• Hard topics: second half

• Best aspect: labs

• I would like to change: “a move away from extensive use of programming”

• “F*%! Matrices”

Statistical Results

• Java with Python better than Java without (p=0.001)

• Java 2009 better than Java 2008 (p=0.0007)

• Java without Python (2009) not better than Java 2008 (p=0.18)

Conclusions

• Python is a joy to teach

• Learning Python is a good pre-cursor for learning Java

• Open source textbooks are fantastic

• Mastery tests are masterful

• Force students to attend labs