Teaching Software Maintenance using 'NET and Rotor

1
Teaching Software Maintenance using .NET and
Rotor

• Leonardo Bottaci
• University of Hull, Hull, UK

2
.NET MSc in Distributed Systems Development

• Aims
• Teach the knowledge and skills to develop and
  maintain large scale systems software.
• Modules
• C
• .NET Framework
• Software maintenance
• Distributed systems programming
• Virtual machine architectures
• Trustworthy computing

3
Software Maintenance Module

• Aims
• The management and practice of software systems
  maintenance
• Difficulties
• No general systematic organised body of
  knowledge, there is no theory.
• Little enthusiasm or glamour, programmers feel
  constrained by legacy and commercial constraints.

4
Student background

• Degree in computer science or equivalent
  experience.
• No student very experienced in
• object oriented languages
• language implementation
• runtime systems
• Relevant for the practical work

5
Software Maintenance Module

• Teaching strategy
• Student centred learning through practical
  experience.
• Laboratory practical work 2 supervised hours per
  week, 1 lecture per week.
• Students may also ask me questions outside those
  times.

6
Software maintenance management 1

• Why manage
• Commercial constraints
• Process models
• IEEE standard
• Matching process models to different types of
  organisation and situations
• Legacy systems
• Safety critical
• Rapid applications development

7
Software maintenance management 2

• Configuration management
• Requirements driven
• Tools
• Code modified in practical work managed using VSS
• Cost estimation
• Formal, organisational level
• Metrics, statistical regression
• Less formal, individual level

8
Software maintenance management 3

• Maintenance through reuse
• All levels, from ideas to code
• Program design
• Program objects should model stable application
  objects
• Functionality is usually not stable
• Seminar discussion of examples from JSD
• .NET contribution to reuse covered in another
  module

9
Maintenance practice

• Course practical work, 100 of credit
• Objective Learn how to maintain software
• Context Modification of the jscript compiler in
  Rotor to produce a program dependency graph
• Emphasis on the learning objective
• Student background varies.

10
Learning Strategy

• Student centred learning through practical
  experience.
• Practice alone reinforces current behaviour.

11
Assessing learning

• In order of importance
• Evidence of learning in the logbook
• Student contribution to seminars and lab
  discussions
• Assessment of modified software
• Safe environment for sensible risk taking

12
Why use Rotor?

• Code contains very few comments
• Code sufficiently readable for students to make
  progress in the relatively short time allocated
  for a module
• Students motivated by real code

13
Initial lab exercises

• Read the Rotor documentation and build the system
• Try the jscript compiler.
• Find, compile and execute a sample jscript
  program
• Modify the jscript compiler by adding a print
  message, rebuild and recompile jscript sample.

14
Maintenance task stage 1

• Modify the jscript compiler to output the
  abstract syntax tree (useful in stage 2 of the
  practical work.)
• Students given a short review of compiler
  operation, scanner, parser, code generator
• No other information

15
Student reaction to task

• Many questions, I answer very few
• The learning more important than the task
• Mantra, logbook
• Some students frustrated, fear of losing marks,
  try to renegotiate task
• Leads to discussion on the role of risk
• need to manage it
• university environment relatively safe.

16
Progress on stage 1

• A few students modify scanner to print characters
  and tokens, boosts confidence
• Students working individually but ideas spread
  fast in the lab
• Students still cannot find the point at which the
  parser returns an abstract syntax tree
• Do not know C
• Insufficient analysis of code, tools
• cordbg used but not effectively

17
Code reading skills

• Code reading should be goal-directed
• Reading to see what is there
• Trying to understand each line
• What are you expecting to find?
• Formulate an hypothesis
• Read the code to confirm or disprove it.

18
Code reading illustration 1

• What is the following code doing?
• while (...)
• ...
• Hypothesise the most popular uses of a loop in
  general and look at code for evidence.

19
Code reading illustration 2

• while (...)
• sum sum ai
• ...
• Array accumulation a likely hypothesis

20
Code reading illustration 3

• while (...)
• sum sum ai
• if (...)
• ...
• else
• ...
• ...
• Hypotheses to explain the conditional inside a
  loop

21
Code reading illustration 4

• while (...)
• sum sum ai
• if (...)
• done 1
• else
• ...
• ...
• Flag, is it for early termination?

22
Code reading illustration 5

• while (i lt 9 and done 0)
• sum sum ai
• if (...)
• done 1
• else
• ...
• ...

23
Maintenance as goal-directed activity

• Debugging should be goal-directed
• Ineffective experience with cordbg
• Students claim they lack knowledge to formulate
  goals
• Then acquisition of knowledge is goal
• Muted student response
• Implications too disturbing

24
AST located, how to print?

• Lack of experience in object-oriented programming
• Mention ToString() after students admit they have
  found no elegant solution
• Still reluctant
• Too ad hoc
• Too many classes to modify
• Place students into teams, do not mix ability.

25
Students recognise lack of knowledge in object
orientation

• Students uncomfortable with lack of design
  knowledge
• Students now told to find out about object
  oriented compilers and object oriented design in
  general
• Use literature, user groups, individuals, code
  itself
• Some students want to learn too much, others too
  little.

26
Design patterns

• Students prompted to research design patterns
• Have just met these in the C module
• Quickly find relevant material on the template
  pattern and visitor pattern

27
Maintenance task stage 2

• Identify through a systematic enumeration the
  occurrences of arithmetic and logical operators
  in a jscript program
• if (i j gt 0)
• j j 1
• else
• i i - 1
• print(i i , j j)

28
Maintenance task stage 2

• Given a program and an operator occurrence id,
  generate a mutant program that differs from the
  given program only at the occurrence of the
  operator
• if (i - j gt 0) //ORIGINALLY i j
• j j 1
• else
• i i - 1
• print(i i , j j)

29
Maintenance planning

• Some students suggest modifying the IL
• Seen as a quick fix
• Two implementation plans
• Modify ast
• Modify IL
• How to evaluate?

30
Detailed plan

• For each proposal
• Plan the implementation steps
• Assumptions that need to be tested
• Decision tree discussion
• Cost estimates in hours

31
Cost estimation individual

• Necessary and frequent activity, usually implicit
• In practical work, students encouraged to make
  cost estimation explicit so that it can be
  scrutinised and improved.
• Calculate estimate, record in logbook
• When task complete, review estimate
• Note how it can be improved

32
Misguided judgement

• Enthusiasm for the IL modification option based
  on difficulty with ast.
• Understanding of requirements conveniently vague.
• Drip feed reminders about the requirements to the
  extent that both options remain in contention.

33
Implementation options evaluated

• IL modification is seen correctly as an easy way
  to solve 95 of the problem
• Remaining 5 hard
• Discussion about the language independent
  generation of mutant programs
• Students return to modification of the ast

34
Practical exercise outcomes

• Learn what is required to maintain software.
• Learn how to improve ones knowledge and skill.
• Lazy practice makes permanent
• Goal directed practice makes better
• Motivation and self confidence.
• Requires a rational assessment of ones abilities

35
Teaching method

• Opportunistic and improvisational
• Respond to issues raised by students
• Easier with small group of students
• Course notes are structured
• Use white board, few slides
• Interact with students to provoke a response

36
Course materials

• www2.dcs.hull.ac.uk/people/cssdjg/MSRAD
• Lecture notes
• Practical exercise
• Tutors notes
• Hints, Modifications to jscript compiler
• Debugging tutorial
• C compiler, MSIL, fjit
• Contact email L.Bottaci at dcs.hull.ac.uk