Sommerville

1
Sommerville

• Chapter 8 System models

2
Kap 8 System models

• Goal
• Document the system's functionality
• Understand the system's functionality
• Communicate with users
• Static system aspects
• Context models
• Data models
• Object models
• Dynamic system aspects
• Behavioural models
• Structured development methods
• CASE tools

3
System models

• About each model type
• What a model type can represent
• Strengths
• Weaknesses
• Different models show different system
  perspectives
• Context model the systems relationship to its
  surroundings
• Behavioural models the system processes (how
  the system behaves) and how the system parts
  work together
• Data models the data / entities in the system
• Architectural models the subsystem structure

4
Model types

• Data processing
• How data is processed at different places the
  system
• Domain
• How entities are composed and related
• Architecture
• System components and the relationships between
  them
• Classification
• A hierarchical organisation of entities with
  common properties
• Stimulus/response
• The system's response to events / messages

5
Context and architectural model

• Context models show the context in which the
  system operates
• System borders
• What is inside and what is outside the system
  borders
• What is covered by the system and what is not
• Architecture models show the system's main
  components and how they cooperate
• Internal components within the system borders
• External components outside the system borders

6
Context model ATM
System border
7
Process model

• Shows the most important system processes
• Main processes
• Supporting processes
• Dataflow models - flow of data between processes
• Input
• Output (result)
• State models - how the system reacts to different
  messages / stimuli
• States
• Stimulus (event)
• State transitions
• Dataflow and state models show different
  perspectives

8
Dataflow model

• Dataflow diagram (DFD) can be used to model how
  data is processed in the system
• A DFD shows
• Data elements
• Data store
• Processes
• How data flows through the system
• Which processing steps are happening in which
  order
• Input to and output from each process
• UML
• Activity diagram

9
Dataflow diagram Procurement
Data element
Process
Data store
10
Dataflow diagram Insulin pump
11
Dataflow diagram - properties

• Models the system's functions
• Showing and documenting the relationships between
  the system data and system processes
• Gives a good understanding of the system
• Can also be used to show dataflow towards
  external modules - outside the system borders
• Used in many different system analysis models
• Simple, intuitive, easy to understand
• Can show data processing from beginning to end

12
State model

• A state model shows how (parts of) the system
  react(s) to internal or external messages /
  stimuli
• A state machine is a graph
• Node state ( action done when the state is
  reached)
• Edge state transition
• Label on edge Stimulus (event) to initiate the
  transition
• An event creates a state transition defined by
  the graph
• A state can have sub-states
• UML
• State diagram

13
State diagram Microwave oven
State
Action
State transition
14
State description - Microwave oven
15
Stimulus description - Microwave oven
16
Substate Microwave oven
17
Semantic data model

• A semantic data model describes the logical
  structure of the data processed by the system
• Frequently used in database design
• Easy to implement as a relational schema
• Most important constructs
• Entity
• Attribute
• Association (relationship)
• Generalisation, aggregation
• Concept definitions - Glossary
• Diagram
• Entity-Relationship (ER) diagram
• UML Class diagram

18
Semantic data model - LIBSYS
19
Glossary

• Definition of all concepts used in the system
  models
• Describes the meaning of each
• Entity
• Attribute
• Association
• Benefits
• Complete definition of all concepts in one place
• Highlights use of duplicate concepts (synonyms)
• Can be used in all parts of the development
  process
• Simplifies training of new personnel

20
Glossary LIBSYS
21
Object oriented models

• Describes a system using classes, operations,
  class cooperation and behaviour
• Extension of the Entity-Relationship model
• Class Entity
• Can model different aspects of data / entities
• Generalisation hierarchy
• Composite data types / Aggregation
• Object behaviour
• Operations (methods)
• Can model different aspect of the system
• Domain model / Design model
• Interaction between objects (user / system /
  classes)

22
Generalisation hierarchy

• Extracting common properties between classes into
  common superclasses
• Can build a hierarchy of super- and subclasses
• Inheritance A subclass inherits alle the
  properties of its superclass
• Attributes
• Operations
• Associations
• UML
• Class diagram
• Inheritance

23
UML Class hierarchy LIBSYS (1)
24
UML Class hierarchy LIBSYS (2)
25
Multiple Inheritance - LIBSYS
26
Aggregation E.g. Course
27
Object behaviour models

• Showing how objects / modules / actors cooperate
  in the system
• Can be used at many levels
• User dialogue
• Communication / cooperation between modules
• Communication / cooperation between classes
  program logic
• UML
• Sequence diagram
• Cooperation diagram
• Activity diagram

28
Sequence diagram User dialogue
29
Structured development models

• Systematic process models for system development
• Describes a process to develop systems
• Guidelines to the contents of a good development
  model
• Many different models have been developed
• Started in the 1970s
• Some important models
• Structured Analysis and Design Method
• SSADM, Yourdon, Jackson (JSD)
• Object Oriented Analyse and Design
• Jacobson (OOSE), Rumbaugh (OMT), Booch
• Unified Process
• Collection of the best elements from many models

30
Structured development models - disadvantages

• Only covers functional requirements not
  non-functional requirements
• Some require producing a lot of documentation
• Models can be hard for users to understand
• A lot of detail
• Difficult notation
• Reaction 1990s
• Agile development methods
• XP

31
CASE tools

• Computer tools to support the development process
• Requirement specification
• Analysis
• Design
• Implementation
• Testing
• Version control
• Reverse engineering
• Often focused towards a particular analysis
  method
• Supplier company - example
• IBM Rational

32
Analysis and design tools
33
System models Summary (1)

• A model is an abstraction describing one aspect
  of a system
• Different models describe different aspects
• Static aspects
• Dynamic aspects

34
System models Summary (2)

• Context and structure models
• How the system relates to its surrounding
  environment external systems and processes
• How the system is composed from (smaller)
  components
• Data models
• Which data are used in the system and how they
  are structured
• Semantic data models constraints and logical
  relationships between data element
• Object models
• entities, classification, aggregation, cooperation

35
System models Summary (3)

• Process models
• Which processes take place in the system
• Dataflow models - how data is flowing through
  and being processed in the system
• State models - how the system reacts to internal
  and external events / stimuli
• Sequence models - Interaction between actors and
  modules in the system
• Structured development models is a framework for
  how to develop system models