Data Structures is traditionally a “bogeyman” discipline in Computer Science courses and has a high degree of failure. In FATEC São José dos Campos we are adopting a hybrid approach, with C and Python languages. The failure rate decreased from 85% (2008) to 10% (2015). The talk will be extensively illustrated with code in C and Python, addressing the various concepts taught in this course. Presented at PyConUS 2016 (Education Summit), PyCon Japan 2015 and EuroPython 2015 (Video: https://youtu.be/6yF5KuOm2LQ?t=226)

1. Data Structures Discipline with Python
@fmasanori

2. I love teaching
CS Professor at FATEC
https://about.me/fmasanori
http://pycursos.com/python-para-zumbis/

6. Difficult with C language
"A C program is like a fast dance on a newly
waxed dance floor by people carrying razors."
Waldi Ravens. Programmer.

7. But I love Data Structures

8. Data Structures are cool

9. Data Structures with Python at FATEC

10. Retention 2008
85% retained
15% approved
Data Structures with C (2008)

11. Retention 2015
10% retained
90%approved
Data Structures with Python (2015)

12. C only:
– 85% retained (2008)
Python + C:
– 12% retained (2014)
– 10% retained (2015)
– ENADE grade 5/5 (max) (last grade)
– 1st Programming Contest InterFATECs
(1st/62) (2014 and 2015)
– Final Brazilian ICPC Programming 2015
Summary

13. • Lab Only – 4 classes/week
• 4 Lab Projects (Python)
• Big Brother (some of the best students could
help the other students as coaches)
• Algorithms in Python + C (side by side)
Details

14. Why Python?
http://cacm.acm.org/blogs/blog-cacm/176450-python-is-now-the-most-popular-
introductory-teaching-language-at-top-us-universities/fulltext

15. Usability is a problem for DS also...
"Results show that many aspects of traditional C-
style syntax, while it has influenced a generation of
programmers, exhibits problems in terms of
usability for novices".
"Perl and Java did not accuracy rates significantly
higher than a language with randomly generated
keywords"
Andreas Stefik and Susanna Siebert: "An Empirical Investigation into
Programming Language Syntax." ACM Transactions on Computing
Education, 13(4), Nov. 2013.

16. The most common fault in computer classes is to
emphasize the rules of specific programming
languages, instead of to emphasize the algorithms
that are being expressed in those languages. D. Knuth
interview at People of ACM, June, 2014.

17. Talk is cheap.
Show me the code.
Linus Torvalds

18. • C is good for optimization (details, low level)
• Python is good to show the essence of the
algorithms (readability, high level)
• If the algorithm is the same (complexity)
“premature optimization is evil” also in
teaching Data Structures
• We are using Python + C with success
TL DR

19. Variables are just names (references)

20. References == “pointers”

21. Big integers

22. Natural integer division

23. Multiple assignment

24. "The programming activity should be viewed
as a process of creating works of literature,
written to be read. “
--D.E. Knuth
Identation

25. Identation in C

26. Identation in C

27. Recursion
"To understand recursion, one must first
understand recursion."
--folklore
"To solve the problem, I found barriers within
barriers. So, I adopted a recursive solution. "
--a student
Ref.: Feofiloff, P., Algoritmos em C, Editora Campus, 2009.

28. Recursion

29. Recursion

30. Recursion

31. Recursion (student solution for dec2bin problem)

32. Linked Lists

33. Linked Lists

34. Linked Lists

35. FIFOs: Distance in Networks

36. FIFOs: Distance in Networks

37. FIFOs: Distance in Networks

38. Stacks: well-formed expression

39. Stacks: well-formed expression

40. Selection Sort

41. Selection Sort

42. Quicksort

43. Quicksort

44. Word Count
Word Count, please download http://www.gutenberg.org/cache/epub/11/pg11.txt

45. Projects: Sorting Algorithm comparison

46. Projects: Arthur Merlin Games simplification
Jessica Pâmela
Leandro
Fernanda Renata
WalberGabrielDiogo EclisBrunoAdriano
Leandro
Walber
Gabriel Diogo
Eclis
Bruno
Adriano

47. Projects: Binaries regions

48. Projects: Minimum Degree Greedy Heuristic for MIS

49. • C is good for optimization (details, low level)
• Python is good to show the essence of the
algorithms (readability, high level)
• If the algorithm is the same (complexity)
“premature optimization is evil” also in
teaching Data Structures.
• We are using Python + C with success
Conclusions

50. Questions?
gist.github.com/fmasanori
http://about.me/fmasanori
fmasanori@gmail.com
Slides: bit.ly/python-DS