Ada%20as%20a%20%20Language

1
Ada as a Language
Teaching

• Ehud Lamm
• The Open University of Israel

2
Context

• Current course Software Engineering with Ada
  (Booch).
• Good, but perhaps not good enough.
• Ada perceived as hindrance.
• I examined several Java textbooks, and discovered
  Lisokvs Program Development in Java.
• What should be our next step?

3
Why How

• Undergraduate SE course (CS students)
• Classic Software Design
• Traditional approach (imperative, OOP)
• Essentials of Software Engineering
• Modularity, Reliability, Readability,
  Flexibility, Software Life Cycle
• University education

4
Why How

• Practical course
• Challenging design problems
• Programming Should be Fun
• Not just getting it to work
• Students evaluate and re-evaluate their designs.
• Make the language invisible

5
Course Outline

• Data Abstraction
• ADT/ADO
• Generic Units
• Inheritance
• Basic Tasking

• Interface/Implementation
• Adequacy
• Operation Categories
• Exceptions
• Pre/Post Conditions
• Invariants

• Flexibility
• Parameterization
• Separate Compilation
• Combining abstractions

• Runtime Polymorphism
• Heterogenic collections
• Classwide/Dispatching
• IS-A (LSP)
• Composition (HAS-A)

6
Q1 Log file

• Log ?
• Gen_Log ?
• EventClass ?
• FilterClass
• WriterClass ?
• Dual Buffering

7
Q2 Priority Queue
Priority is priv. with func lt(..)
Priority is (ltgt)
array of queues
sorted linked list
8
generic type Struct is lim. priv. type Elem
is priv. with proc Insert with proc
Remove package Any_Struct is end
Struct is new Any_Struct(ltgt)
Package Protect_Struct protected type
Struct . end
9
pack BST --Binary_Search_Tree
func BST.Max..
10
generic type Item is priv. package Sets is
type Set is abs. tagged null rec end
generic with func lt(I1,I2Item) package
Sets.Sorted_List is type Set is new Sets.Set
with priv.
11
Does the choice of programming language really
matter?
12
No

• but...

13
In theory, theres no difference between theory
and practice In practice, there is.
14
Ideal Language
Ada
Java

• Rich type system
• Information hiding
• Explicit interfaces
• Value semantics
• Genericity
• Inheritance
• Garbage collection
• Design by Contract
• First-class functions
• Rich libraries, variety of tools

15
Language Community

• Libraries
• Reusable code
• Tools (e.g., VS IDEs, diagramming, DbC, etc.)
• Tutorials/textbooks
• etc.

16
Problematic features (Ada)(in my experience)

• Many kinds of Interfaces
• Access to subprograms (accessibility)
• Controlled (vs. other approaches to
  destructors/constructors)
• Controlled Generics
• Controlling child visibility
• Verbosity (exercises take too much time)

17
Problematic features (Java)

• Conjecture. Let me know what you think.
• The type system
• Reference semantics
• Low level concurrency
• Generics are coming..

18
The textbooks

• A language cant survive without good textbook,
  for all levels and needs.
• Liskovs book is much more suited for an academic
  SE course.

19
Conclusions

• Ada is still a strong candidate for a teaching
  language for SE
• Most of the technical problems can be solved
  rather easily.
• Better error messages should help with the
  remaining issues.
• The language isnt the problem Good News and Bad
  News
• Community and Cultural roadblocks

20
Conclusions

• Can we and should we try to change the way Ada is
  perceived in universities?
• How?
• Evangelism
• Textbooks
• GNAT
• etc.

21
Thank You
Any Questions?