1. On Teaching Introductory Programming Abhiram Ranade

2. Background I recently taught the IITB introductory programming course, CS 101. –700 students, varying preparation –60 Teaching Assistants –Lack of a good textbook –All usual teaching related issues: content, delivery. Report of personal experience. –Main focus: content and delivery

3. Outline The new student Basic Premises on learning: motivation, fun. Course Themes: –Graphics –Every lecture must solve a nice problem –Depth helps learning –Connect to other disciplines –Cool Algorithms/Fun Concluding remarks

4. The New Student Less reverent than 20 years ago More oppressed in school years, expects more freedom in college. Wants to know why he/she should learn what is being taught Is swayed by immediate gratification. Has many distractions

5. Basic Premises on learning Motivation must be very clear –The student must develop ownership of the questions being answered in the course Topics must be fun! Beauty vs. Utility? Beauty and utility.

6. Fun Eating ice cream Climbing a mountain Swimming or playing a sport Seeing a movie, natural panorama Seeing a magic show Learning should have all kinds of fun!

7. Fun and programming Sense of accomplishment: students should be able to do something significant of interest to themselves. Should relate CS to other sciences, even life! Should be surprised by cool algorithms. Must be occasions of “Wow!”.

8. Theme 1: Graphics + Turtle Geometry Available to students, from lecture 1. Used to teach many concepts: –Notion of program as sequence of instructions. –Iteration: repeated patterns –Recursion: recursive patterns –Procedures: draw_house(); –Graphical debugging –Visualization: Trees, Newton’s method for finding roots.

9. Turtle Geometry + Graphics (contd) Class heirarchy: Shape Line PolygonCircle Turtle

10. Introductory Lecture Traditional: Printing “Hello world” –Lots of syntax with little excitement. Instead: use graphics package to draw nice pictures. Non trivial picture drawing exercises from day 1. Lecture 1 sets the tone of the course.

11. Theme 2: Begin with a problem Every lecture begins with a problem:e g. we want to plot movement of gas particles on the screen. Can we do this using the language we know so far? No. Here is a new statement/concept which will help. Not: “Today we will learn about arrays..”

12. Theme 3: Depth helps learning For every concept have –good motivating examples –discuss implementation, e.g. recursion, pointers –interesting programming exercises. Generalization helps learning: –pointers: lists are good, but trees really force you to learn! Mix ideas: Use graphics to draw out trees recursively...

13. Connect to Other Disciplines Newton’s method Simulate movement of charged particles Hemachandra (Fibonacci) numbers as they arise in counting the number of different possible poetic meters of n beats.

14. Cool Algorithms Newton’s algorithm Searching trees: visualize with turtle moving on the screen. Shortest paths: good combination of –matrix multiplication over +, min –C++ supports representation of infinity: “HUGE_VAL” ==> very elegant, short code for computing shortest paths –Students feel they are improving their math!

15. Fun! Graphics package includes X windows “event loop”: wait for event to happen and based on event do whatever. Used by many students to develop “video games” in final project Did not talk about event driven programming, listeners etc. because too much syntax.

16. General Remarks Carefully designed lab exercises. TAs graded them. Final weightage of labs was less because of fear of copying. Students liked assignments and said they helped learn. Switched to writing on board (tablet) because of student demand. Students said lecture developed better than slides. Some students found the course too mathematical.

17. What I would have liked to do but didnt String class examples Assertions, reasoning about programs.. STL and algorithms class use. Programming style Debuggers/IDEs. Would have been possible with 3 lectures/week.