Video and slides synchronized, mp3 and slide download available at http://bit.ly/13YbQvb. Abstraction is at the center of much work in Computer Science. It encompasses finding the right interface for a system as well as finding an effective design for a system implementation. Filmed at qconlondon.com. Barbara Liskov is an Institute Professor at MIT, a member of the National Academy of Engineering and the National Academy of Sciences, a fellow of the American Academy of Arts and Sciences, and a fellow of the ACM. Awards received: ACM Turing Award, the ACM SIGPLAN Programming Language Achievement Award, the IEEE Von Neumann medal, lifetime achievement award from the Society of Women Engineers.

1. The Power of Abstraction
Barbara Liskov
March 2013
MIT CSAIL

2. InfoQ.com: News & Community Site
• 750,000 unique visitors/month
• Published in 4 languages (English, Chinese, Japanese and Brazilian
Portuguese)
• Post content from our QCon conferences
• News 15-20 / week
• Articles 3-4 / week
• Presentations (videos) 12-15 / week
• Interviews 2-3 / week
• Books 1 / month
Watch the video with slide
synchronization on InfoQ.com!
http://www.infoq.com/presentations
/programming-abstraction-liskov

3. Presented at QCon London
www.qconlondon.com
Purpose of QCon
- to empower software development by facilitating the spread of
knowledge and innovation
Strategy
- practitioner-driven conference designed for YOU: influencers of
change and innovation in your teams
- speakers and topics driving the evolution and innovation
- connecting and catalyzing the influencers and innovators
Highlights
- attended by more than 12,000 delegates since 2007
- held in 9 cities worldwide

4. Software is Complex
 Systems are big
 and they do complicated things
 and they may be distributed and/or
concurrent

5. Addressing Complexity
 Algorithms, data structures, protocols

6. Addressing Complexity
 Algorithms, data structures, protocols
 Programming methodology
 Programming languages

7. This Talk
 Programming methodology as it
developed
 Programming languages
 Programming languages today

8. The Situation in 1970
 The software crisis!

9. Programming Methodology
 How should programs be designed?
 How should programs be structured?

10. The Landscape
 E. W. Dijkstra. Go To Statement
Considered Harmful. Cacm, Mar. 1968

11. The Landscape
 N. Wirth. Program Development by
Stepwise Refinement. Cacm, April 1971

12. The Landscape
 D. L. Parnas. Information Distribution
Aspects of Design Methodology. IFIP
Congress, 1971
 “The connections between modules are
the assumptions which the modules
make about each other.”

13. Modularity
 A program is a collection of modules

14. Modularity
 A program is a collection of modules
 Each module has an interface,
described by a specification

15. Modularity
 A program is a collection of modules
 Each has an interface, described by a
specification
 A module’s implementation is correct if it
meets the specification
 A using module depends only on the
specification

16. Modularity
 A program is a collection of modules
 Each has an interface, described by a
specification
 A module’s implementation is correct if it
meets the specification
 A using module depends only on the
specification
 E.g. a sort routine sort(a)

17. Benefits of Modularity
 Local reasoning
 Modifiability
 Independent development

18. The Situation in 1970
 Procedures were the only type of
module
 Not powerful enough, e.g., a file system
 Not used very much
 Complicated connections

19. Partitions
 B. Liskov. A Design Methodology for
Reliable Software Systems. FJCC, Dec.
1972

20. Partitions
Partition state
op1 op2 op3

21. From Partitions to ADTs
 How can these ideas be applied to
building programs?

22. Idea
 Connect partitions to data types

23. Meeting in Savanah
 ACM Sigplan-Sigops interface meeting.
April 1973. (Sigplan Notices, Sept.
1973)
 Started to work with Steve Zilles

24. The Landscape
 Extensible Languages
 S. Schuman and P. Jourrand. Definition
Mechanisms in Extensible Programming
Languages. AFIPS. 1970
 R. Balzer. Dataless Programming. AFIPS.
1967

25. The Landscape
 O-J. Dahl and C.A.R. Hoare. Hierarchical
Program Structures. Structured
Programming, Academic Press, 1972

26. The Landscape
 J. H. Morris. Protection in Programming
Languages. Cacm. Jan. 1973

27. Abstract Data Types
 B. Liskov and S. Zilles. Programming
with Abstract Data Types. ACM Sigplan
Conference on Very High Level
Languages. April 1974

28. What that paper proposed
 Abstract data types
 A set of operations
 And a set of objects
 The operations provide the only way to use
the objects
 A sketch of a programming language

29. From ADTs to CLU
 Participants
 Russ Atkinson
 Craig Schaffert
 Alan Snyder

31. Why a Programming
Language?
 Communicating to programmers
 Do ADTs work in practice?
 Getting a precise definition
 Achieving reasonable performance

32. Some Facts about CLU
 Static type checking
 Heap-based
 Separate compilation
 No concurrency, no gotos, no
inheritance

33. CLU Mechanisms
 Clusters
 Polymorphism
 Iterators
 Exception handling

34. Clusters
IntSet = cluster is create, insert, delete, …
% representation for IntSet objects
% implementation of the operations
end IntSet

35. Clusters
IntSet = cluster is create, insert, delete, …
% representation for IntSet objects
% implementation of the operations
end IntSet
IntSet s = IntSet$create( )
IntSet$insert(s, 3)

36. Polymorphism
Set = cluster[T: type] is create, insert, …
% representation for Set object
% implementation of Set operations
end Set
Set[int] s := Set[int]$create( )
Set[int]$insert(s, 3)

37. Polymorphism
Set = cluster[T: type] is create, insert, …
where T has equal: proctype(T, T)
returns (bool)

38. Iterators
 For all x in C do S

39. Iterators
 For all x in C do S
 Destroy the collection?
 Complicate the abstraction?

40. Visit to CMU
 Bill Wulf and Mary Shaw, Alphard
 Generators

41. Iterators
sum: int := 0
for e: int in Set[int]$members(s) do
sum := sum + e
end

42. Also
 Exception handling
 Strong specifications, e.g., IntSet$choose
 First class Procedures and Iterators

43. After CLU
 Argus and distributed computing
 Programming methodology
 Modular program design
 Reasoning about correctness
 Type hierarchy

44. From CLU to Object-Oriented
Programming
 SmallTalk provided inheritance

45. The Landscape
 Inheritance was used for:
 Implementation
 Type hierarchy

46. Type Hierarchy
 Wasn’t well understood
 E.g., stacks vs. queues

47. The Liskov Substitution
Principle (LSP)
 Objects of subtypes should behave like
those of supertypes if used via
supertype methods
 B. Liskov. Data abstraction and
hierarchy. Sigplan notices, May 1988

48. What Next?
 Modularity based on abstraction is the
way things are done

49. Programming Languages
Today
 Languages for experts, e.g., Java, C#

50. Programming 1A
 E.g., Python

51. Challenges
 A programming language for novices
and experts
 Ease of use vs. expressive power
 Readability vs. writeability
 Modularity and encapsulation
 Powerful abstraction mechanisms
 State matters

52. Challenges
 Massively-parallel computers
 Programming methodology
 Programming language support

53. The Power of Abstraction
Barbara Liskov
March 2013
MIT CSAIL