CP3015 Rapid Application Development

1
CP3015 Rapid Application Development

• An introduction to some of the background
  concepts

2
RAD is called a methodology

• but what is a methodology, why are they
  developed, how do you tell a good one, what do
  they have in common ?

3
A methodology is ......

• a pompous word for method
• investigation, research and discourse about
  methods
• a set of techniques, tools, methods with some
  underlying philosophy

4
The last definition is probably what we need here

• So we have to look for philosophies as well as
  tools, methods etc

5
Some methodologies ...... with examples of
specific methods

• Structured methodologies
• Yourdon, SSDAM, Ward-Mellor
• Object-oriented methodologies
• OMT, Shlaer-Mellor
• Soft systems methodologies
• Ethics, Multiview
• Formal methodologies
• Z, VDM, CSP, Z

6
Some Tools

• ERDs
• DFDs
• Rich pictures
• STDs
• Structure charts
• Formal logics
• Object models

• Prototypes
• ELHs
• JSP
• Normalisation
• Data dictionaries
• PDL
• Decision tables / trees

7
Look up any

• that youre not familiar with !

8
Each methodology will use several of these tools

• with different emphasis, for different reasons,
  in a different order...

9
New methodologies evolves because

• perceived weaknesses in all other methodologies
• changed and improved technologies
• case, programming languages, hardware
• changed environments, markets and needs
• trends, fashion, band-wagon, legislation etc

10
A quick run through some methodologies (and
methods)

• What led to their rise, then to their fall what
  were their tools and philosophies ?

11
Structured Methods

• Probably the first methodology, after the
  so-called Test-and-Build i.e. hacking

12
Structured Methods

• Structured analysis / design / programming
• started in large-scale, stable, centralised
  corporate DP on mainframes using COBOL
• waterfall life cycle
• top-down / functional decomposition
• tool-kit vs cook-book
• e.g. Yourdon vs SSADM

13
Structured Methods Tools

• DFDs
• Structure Charts
• ERDs
• Data dictionaries

14
Extensions to Structured Methods

• Real-time Ward-Mellor, Pirbai-Hatley
• STDs, events, control flow
• Requirements Capture
• prototyping added
• SSADM v4
• adds SSM

15
Problems with Structured Methods

• Paralysis-by-analysis / Victorian novel
• Increasingly unstable requirements they change
  quicker and quicker
• validation customers / users dont understand
  DFDs, ERDs etc
• verification lurch from one notation to the next

16
Problems with Structured Methods

• lack of precision, completeness, consistency etc
  much natural language is used in Structured
  Methods
• lack of user involvement
• feasibility sometimes poorly identified
• may not scale up or down

17
Formal Methods

• Using maths

18
More Problems with Structured Methods

• Structured methods merely turn a big bit of vague
  text into small bits of vague text embedded in an
  ambiguous DFD
• Even if the structured analysis is provably
  correct, doesnt mean the design and source code
  will be ! Or the machine code !

19
Formal Methods Tools

• propositional and predicate calculus
• discrete maths and logic

20
Formal Methods consist of

• Formal specifications
• the specification expressed in predicate calculus
• animated with tools and checked e.g. Cadiz
• some are executable e.g. Metoo, Prolog
• difficult

21
Formal Methods consist of

• Formal development
• provable progression to code
• very difficult

22
Formal Methods problems

• Too expensive, probably too academic
• customers not convinced, developers not fond.
• Only deals with one set of problems
• Inadequate tool support
• Maybe some future in safety- / mission- / life-
  critical systems

23
Soft Systems

• One response to the failings of Structured Methods

24
Origins

• Structured methods deal with problems
  organisations have difficulties
• Most information processing systems
• have substantial human / social components
• are embedded in social organisations
• Customers and users have partial and conflicted
  views on their problems

25
Soft Systems Tools

• Based on systems theory
• holistic not reductionist
• subjective not objective
• fuzzy not well-defined
• purpose, boundaries, components
• iterative

26
Soft Systems Tools

• Root definition
• precise verbal description of the essence of the
  processes
• CATWOE
• Rich pictures
• to build and check a conceptual model that form
  basis of information systems

27
Soft Systems

• Can be
• grafted onto front-end of hard systems methods
• have hard systems methods embedded within them
• are used increasingly

28
Object-Oriented Methods

• Another response to the failings of Structured
  Methods

29
Object-Oriented Methods

• Grew out of programming
• eg Smalltalk, Simula
• Originally technical and academic
• Intuitive world consists of things, not
  functions
• Offers framework for re-use
• hence quicker quality
• Now have industrial-strength methods

30
Object-Oriented Methods

• Assume solution-domain will be entirely
  computer-based
• Use one notation throughout life-cycle

31
Object-Oriented Tools

• Object models
• an enhanced ERD
• Dynamic model
• Harel charts
• Process model
• DFDs