Site MapHelpFeedbackChapter Summary
Chapter Summary
(See related pages)

Overview

The purpose of component level design is to define data structures, algorithms, interface characteristics, and communication mechanisms for each software component identified in the architectural design. Component level design occurs after the data, architectural, and interface designs are established. The component-level design represents the software in a way that allows the designer to review it for correctness and consistency, before it is built. The work product produced is a design for each software component, represented using graphical, tabular, or text-based notation. Design walkthroughs are conducted to determine correctness of the data transformation or control transformation allocated to each component during earlier design steps.

Component Definitions
  • A component is a modular, deployable, replaceable part of a system that encapsulates implementation and exposes a set of interfaces
  • Object-oriented view is that a component contains a set of collaborating classes
  • Traditional view is that a component (or module) resides in the software and serves one of three roles
    • Control components coordinate invocation of all other problem domain components
    • Problem domain components implement a function required by the customer
    • Infrastructure components are responsible for functions needed to support the processing required in a domain application
  • Process-Related view emphasizes building systems out of existing components chosen from a catalog of reusable components as a means of populating the architecture
Object-Oriented Component Design
  • Focuses on the elaboration of domain specific analysis classes and the definition of infrastructure classes
  • Detailed description of class attributes, operations, and interfaces is required prior to beginning construction activities
Principles for Designing Class-based Components
  • Open-Closed Principle (OCP) - class should be open for extension but closed for modification
  • Liskov Substitution Principle (LSP) - subclasses should be substitutable for their base classes
  • Dependency Inversion Principle (DIP) - depend on abstractions, do not depend on concretions
  • Interface Segregation Principle (ISP) - many client specific interfaces are better than one general purpose interface
  • Release Reuse Equivalency Principle (REP) - the granule of reuse is the granule of release
  • Common Closure Principle (CCP) - classes that change together belong together
  • Common Reuse Principle (CRP) - classes that can't be used together should not be grouped together
Component-Level Design Guidelines
  • Components
    • Establish naming conventions during architectural modeling
    • Architectural component names should have meaning to stakeholders
    • Infrastructure component names should reflect implementation specific meanings
    • Use of UML stereotypes may help identify the nature of components
  • Interfaces
    • Use lollipop representation rather than formal UML box and arrow notation
    • For consistency interfaces should flow from the left-hand side of the component box
    • Show only the interfaces relevant to the component under construction
  • Dependencies
    • For improved readability model dependencies from left to right and inheritance from bottom (derived classes) to top (base classes)
    • Component interdependencies should be represented by interfaces rather that component to component dependencies
Cohesion
  • Utility cohesion - components grouped within the same category but are otherwise unrelated
  • Temporal cohesion - operations are performed to reflect a specific behavior or state
  • Procedural cohesion - components grouped to allow one be invoked immediately after the preceding one was invoked with or without passing data
  • Communicational cohesion - operations required same data are grouped in same class
  • Sequential cohesion - components grouped to allow input to be passed from first to second and so on
  • Layer cohesion - exhibited by package components when a higher level layer accesses the services of a lower layer, but lower level layers do not access higher level layer services
  • Functional cohesion - module performs one and only one function
Coupling
  • Data coupling - occurs when long strings of arguments are passed between components
  • Stamp coupling - occurs when parts of larger data structures are passed between components
  • Control coupling - occurs when one component passes control flags as arguments to another
  • External coupling - occurs when a component communicates or collaborates with in infrastructure component (e.g., database)
  • Common coupling - occurs when several components make use of a global variable
  • Content coupling - occurs when one component surreptitiously modifies internal data in another component
  • Routine call coupling - occurs when one operator invokes another
  • Type use coupling - occurs when one component uses a data type defined in another
  • Inclusion or import coupling - occurs when one component imports a package or uses the content of another
Conducting Component Level Design
  1. Identify all design classes that correspond to the problem domain.
  2. Identify all design classes that correspond to the infrastructure domain.
  3. Elaborate all design classes that are not acquired as reusable components.
    1. Specify message details when classes or components collaborate.
    2. Identify appropriate interfaces for each component.
    3. Elaborate attributes and define data types and data structures required to implement them.
    4. Describe processing flow within each operation in detail.
  4. Identify persistent data sources (databases and files) and identify the classes required to manage them.
  5. Develop and elaborate behavioral representations for each class or component.
  6. Elaborate deployment diagrams to provide additional implementation detail.
  7. Factor every component-level diagram representation and consider alternatives.
Object Constraint Language (OCL)
  • Complements UML by allowing a software engineer to use formal grammar and syntax to construct unambiguous statements about design model element
  • Parts of OCL language statements
    1. a context that denotes the limited situation in which the statement is valid
    2. a property representing some characteristics of the context
    3. an operation that manipulates or qualifies a property
    4. keywords used to specify conditional expressions
Conventional or Structured Component Design
  • Each block of code has a single entry at the top
  • Each block of code has a single exit at the bottom
  • Only three control structures are required: sequence, condition (if-then-else), and repetition (looping)
  • Reduces program complexity by enhancing readability, testability, and maintainability
Design Notation
  • Flowcharts - arrows for flow of control, diamonds for decisions, rectangles for processes
  • Decision table - subsets of system conditions and actions are associated with each other to define the rules for processing inputs and events
  • Program Design Language (PDL) - structured English or pseudocode used to describe processing details
Program Design Language Characteristics
  • Fixed syntax with keywords providing for representation of all structured constructs, data declarations, and module definitions
  • Free syntax of natural language for describing processing features
  • Data declaration facilities for simple and complex data structures
  • Subprogram definition and invocation facilities
Design Notation Assessment Criteria
  • Modularity - notation supports development of modular software
  • Overall simplicity - easy to learn, easy to use, easy to write
  • Ease of editing - easy to modify design representation when changes are necessary
  • Machine readability - notation can be input directly into a computer-based development system
  • Maintainability - maintenance of the configuration usually involves maintenance of the procedural design representation
  • Structure enforcement - enforces the use of structured programming constructs
  • Automatic processing - allows the designer to verify the correctness and quality of the design
  • Data representation - ability to represent local and global data directly
  • Logic verification - automatic logic verification improves testing adequacy
  • Easily converted to program source code (makes code generation quicker)







PressmanOnline Learning Center

Home > Chapter 11 > Chapter Summary