Chair of Software Engineering. The 26 concepts. Bertrand Meyer

Transcription

1 Chair of Software Engineering The 26 concepts of computer science Bertrand Meyer Informatics Education Europe Venice, December

2 Five principles 1. We are not in Kansas any more 2. Most students come to us with considerable baggage 3. To teach software, show software 4. The only meaningful test is learning outcomes, objectively assessed 5. To teach CS is to teach 26 concepts 2 2

3 Introductory programming teaching Teaching first-year programming is a politically sensitive area, as you must contend not only with your students but also with an intimidating second audience colleagues who teach in subsequent semesters. Academics who teach introductory programming are placed under enormous pressure by colleagues. As surely as farmers complain about the weather, computing academics will complain about students programming abilities. Raymond Lister: After the Gold Rush: Toward Sustainable Scholarship in Computing, 10th Conf. on Australasian Computing Education,

4 Teaching programming today Ups and downs of high-tech economy, image of CS Offshoring and globalization raise the stakes Short-term pressures (e.g. families), IT industry fads Bologna process Widely diverse student motivations, skills, experience 4 4

5 Student backgrounds, skills The Youtube generation (or Flickr, Second Life, MySpace, Facebook, Grand Theft Auto etc.) Grew up with the Web, video games, multimedia Some have extensive programming experience Others have no clue! 5 5

6 Background of 1 st -year ETH CS students Previous years Computer 1 year: 0% (0%, 0%, 1%) experience 2 to 4 years: 1% (3%, 4%, 1%, 6%) 5 to 9 years: 38% (35%, 48%, 35%, 55%) 10 years: 61% (62%, 48%, 63%) Programming experience None: 12% (19%, 18%, 14%, 22%) No O-O: 20% (26%, 33%, 38%) 100 classes: 8% (11%, 15%, 10%, 5%) Pedroni, Meyer, Oriol, They know more than we think!,

7 Ways to teach introductory programming 1. Programming in the small 2. Learn APIs 3. Teach a programming language: Java, C++ 4. Functional programming 5. Completely formal, don t touch a computer 7 7

8 1. Programming in the small Necessary, but not sufficient This is not the 70 s any more 8 8

9 2. Teach APIs Example: graphical programming using Swing Pros: Focus on abstraction Yields directly employable students But: no understanding of underlying concepts 9 9

10 3. Teaching a specific language First Java program: class First { public static void main(string args[]) { System.out.println("Hello ll World!"); } } You ll understand d when you grow up! Do as I say, not as I do 10 10

11 Clarity Explain all that you show 11 11

12 3. Teaching a specific language First Java program: class First { public static void main(string args[]) { System.out.println("Hello ll World!"); } } 12 12

13 From Graham Farr s slides, intro course, Monash University (Melbourne, Australia) Common Mistakes in while while (num < minimum) scanf( %d, &num); printf( Number must be greater than %d.\n, minimum); printf( Please try again.\n ); while (num < minimum) { scanf( %d, %d,&num); printf( Number must be greater than %d.\n, minimum); printf( Please try again.\n ); } 13 13

14 From Farr s slides (continued) Common Mistakes in while -- extra semi-colon; Marks the end of the while-block -- usual cause of infinite loops while (num < minimum); { scanf( %d, &num); printf( Number must be greater than %d.\n, minimum); printf( Please try again.\n ); } 14 14

15 4. Functional languages Emphasize mathematical nature of programming Abelson-Sussman, MIT course: Structure & Interpretation of Computer Programs Elegant, concentrates on essential concepts But Little connection with industry practice Risk of disconnecting from students Why is state bad? 15 15

16 5. Formal methods from the start [In American universities] there is a complete separation between mathematics and engineering. I know of one highly regarded American university in which students in their first programming course must prove the correctness of every tiny program they write. In their second programming course, mathematics is completely forgotten and they just learn how to write C programs. There is no attempt to apply what they learned in the first course to the writing of real programs. Leslie Lamport (2003) 16 16

17 Concepts or skills? Skills supported by concepts 17 17

18 The twenty-six concepts of computer science Structure Static vs dynamic Notation Reuse Change Abstraction, Recursive reasoning Classification Syntax vs semantics Contract Information hiding Specification vs implementation Algorithmic reasoning Typing Complexity Scaling up Invariant 18 18

19 Computing the maximum of an array from until loop end i := 0 ; Result := a [1] i > a.upper i := i + 1 Result := max (Result, a [i ]) 19 19

20 Levenshtein distance Beethoven to Beatles B E E T H O V E N A L S Operation R D R D R Distance

21 B E A T L E S B E E T H

22 Levenshtein distance algorithm distance (source, target: STRING): INTEGER -- Minimum number of operations to turn source into target locall dist : ARRAY_2 [INTEGER] i, j, del, ins, subst : INTEGER do create dist.make (source.count, target.count) from i := 0 until i > source.count loop dist [i, 0] := i ; i := i + 1 end from j := 0 until j > target.count loop dist [0, j ] := j ; j := j + 1 end -- (Continued) 22 22

23 Levenshtein distance algorithm from i := 1 until i > source.count loop from j := 1 until j > target.count invariant loop??? if source [i ] = target [ j ] then dist [i, j ] := dist [ i -1, j -1] else deletion := dist [i -1, j ] insertion := dist [i, j -1] substitution := dist [i -1, j -1] dist [i, j ] := minimum (deletion, insertion, substitution) + 1 end j := j + 1 end i := i + 1 end Result := dist (source.count, target.count) end 23 23

24 Reversing a list right i first_element pivot from pivot := first_element first_element := Void until pivot = Void loop i := first_element end first_element := pivot pivot := pivot.right first_element.put_right (i ) 24 24

25 Reversing a list right i first_element pivot from pivot := first_element first_element := Void until pivot = Void loop i := first_element end first_element := pivot pivot := pivot.right first_element.put_right (i ) 25 25

26 Reversing a list right i first_element pivot from pivot := first_element first_element := Void until pivot = Void loop i := first_element end first_element := pivot pivot := pivot.right first_element.put_right (i ) 26 26

27 Reversing a list right i first_element pivot from pivot := first_element first_element := Void until pivot = Void loop i := first_element end first_element := pivot pivot := pivot.right first_element.put_right (i ) 27 27

28 Reversing a list right i first_element pivot from pivot := first_element first_element := Void until pivot = Void loop i := first_element end first_element := pivot pivot := pivot.right first_element.put_right (i ) 28 28

29 Loop as approximation strategy a 1 a 2 a i a n Result = a 1 Result = Max (a 1.. a 2 ) Loop body: = Max (a 1.. a 1 ) i := i + 1 Result = Max (a 1.. a i ) Result := max (Result, a [i ]) The loop invariant Result = Max (a 1.. a n ) 29 29

30 Loops as problem-solving strategy A loop invariant is a property that: Is easy to establish initially (even to cover a trivial part of the data) Is easy to extend to cover a bigger part If covering all data, gives the desired result! 30 30

31 B E A T L E S B Invariant: each a [i, j ] is distance from source [1..i ] to target [1..j ] E E T 4 4 K R D I 2 H 5 K Keep I Insert D Delete R Replace 31 31

32 Levenshtein loop from i := 1 until i > source.count loop from j := 1 until j > target.count invariant -- For all p : 1.. i, i q : 1.. j 1, we can turn source [1.. p ] -- into target [1.. q ] in dist [p, q ] operations loop if source [i ] = target [ j ] then new := dist [ i -1, j -1] else deletion := dist [i -1, j ] insertion := dist [i, j -1] substitution := dist [i -1, j -1] new := deletion.min (insertion.min (substitution)) + 1 end dist [i, j ] := new j := j + 1 end i := i + 1 end Result := dist (source.count, target.count) 32 32

33 Reversing a list right i first_element pivot Invariant: from first_element following right, initial i i items in inverse order; from pivot, rest of items in original order from pivot := first_element first_element := Void until pivot = Void loop i := first_element first_element := pivot pivot := pivot.right first_element.put_right (i ) end 33 33

34 Source: Graham Farr, intro course, Monash Common Mistakes in while while (num < minimum) scanf( %d, &num); printf( Number must be greater than %d.\n, minimum); printf( Please try again.\n ); while (num < minimum) { scanf( %d, &num); printf( Number must be greater than %d.\n, minimum); printf( Please try again.\n ); } 34 34

35 The twenty-six concepts of computer science Structure Static vs dynamic Notation Reuse Change Abstraction, Recursive reasoning Classification Syntax vs semantics Contract Information hiding Specification vs implementation Algorithmic reasoning Typing Complexity Scaling up Invariant 35 35

36 Concepts or skills? Skills supported by concepts 36 36

37 Some topics Control structures Undecidability Syntax, BNF Recursion Algorithm complexity Genericity, inheritance (single & multiple), Visitor Polymorphism, dynamic binding Agents (behavior as objects) Event-driven programming, Observer In-depth example: topological sort, problem solving strategies, engineering the solution Introduction to software engineering 37 37

38 Principles of our course Fully object-oriented from the start, using Eiffel Outside-in ( Inverted Curriculum ) Gentle introduction to formal techniques: Design by Contract Textbook: Touch of Class 38 38

39 Outside-in The course gives students a large amount of software, right from the beginning Michela Pedroni & numerous students TRAFFIC library 150,000 lines of Eiffel They start out as consumers They end up as producers! Progressive opening of the black boxes TRAFFIC is graphical, multimedia, extendible, and fun! 39 39

40 The first program class PREVIEW inherit TOURISM feature explore -- Prepare & animate route do Paris.display Louvre.spotlight Metro.highlight end end Route1.animate Text to input 40 40

41 3. Teaching a specific language First Java program: class First { public static void main(string args[]) { System.out.println("Hello ll World!"); } } You ll understand d when you grow up! Do as I say, not as I do 41 41

42 The first program class PREVIEW inherit TOURISM feature explore -- Prepare & animate route do Paris.display Louvre.spotlight Metro.highlight end end Route1.animate Text to input 42 42

43 43 43

44 Principles of the ETH course Reuse software : inspiration, imitation, abstraction See lots of software Learn to reuse through interfaces and contracts Gentle introduction to formal methods Interesting examples from day one Combination of principles and practices All relevant traditional techniques included as well: algorithms, control structures, basic data structures, recursion 44 44

45 To know more touch.ethz.ch (Still for a couple of weeks) including PDF of textbook Springer-Verlag, February

46 The next step Evaluation Second year (using Software Architecture course) Programming Ability Test, based on neutral criteria 46 46

47 Other relevant work Concurrency (SCOOP model) EVE (Eiffel Verification Environment) Automatic testing: AutoTest, CDD Proofs Teaching: DOSE course (Distributed and Outsourced Software Engineering), joint student project with Politecnico di Milano, Univ. of Nijny Novgorod, Odessa Polytechnic, University of Debrecen 47 47

48 Five principles 1. We are not in Kansas any more 2. Most students come to us with considerable baggage 3. To teach software, show software 4. The only meaningful test is learning outcomes, objectively assessed 5. To teach CS is to teach 26 concepts 48 48