LIFE CYCLE OF SOFTWARE ENGINEERING
                   ==================================

References:
  Software Engineering: A Practitioner's Approach by Pressman.
  Software Engineering by Sommerville.
  An Integrated Approach to Software Engineering by Jalote.
  Fundamentals of Software Engineering by Ghezzi, Jazayeri, Mandrioli.
  Classical and Object-Oriented Software Engineering by Schach.

Life Cycle: 
  Often viewed as a "waterfall" from a high-level of abstraction to a low level of      
  abstraction. Sometimes the process must move back to start over again at an earlier 
  phase. The software may follow the classical (structured, function-oriented) or
  object-oriented methodology.

  Most of the following phases have a review and sometimes a set of milestones to check
  whether the process is working appropriately.

  Note in the following that "outputs" are "deliverables".

Planning:
  Goal  : On-going process of planning and organizing resources to achieve effective 
          software development.
  Output: Software Project Plan

Requirements: 
  Goal  : High-level description of the product.
  Output: Requirements Specification Document:
            Goals, Narrative, Priorities, Constraints, System Inputs/Outputs.

Analysis: 
  Goal  : Investigation of "what" the system will do.
  Types : Structured Analysis with Data Flow Diagrams.
          Object-Oriented Analysis with classes and associations.
  Input : Requirements Specification Document
  Output: Analysis Specification Document:
            Models which abstract the system in order to reduce complexity.
            Validation Test Plan
            Optional Prototype for users to test drive the system.
            Preliminary User's Manual.
 
Design: 
  Goal  : Specification for "how" the system will be built.
  Types : Structured Design with transform analysis to produce a control hierachy.
          Object-Oriented Design with objects, attributes, methods.
  Input : Analysis Specification Document
  Output: Design Specification Document
 
    Preliminary Design: 
      
      Data Design: High-level presentation of data structures.
      
      Architecture Design: Structure/Organization of modules or objects.

    Detail Design:

      Detail Data Design: Low-level presentation of data structures.
      
      Procedural Design: pseudocode, flowcharts, or collaboration diagrams.

Implementation:
  Goal  : Effective translation of a design to a programming language.
  Input : Design Specification Document.
  Output: Code.
 
Test:
  Goal  : Systematically uncover errors within a reasonable amount of time and effort.
  Inputs: 
    Project Plan from Planning. 
    Analysis Specification Document including Validation Test Plan. 
    Code from Implementation.
    Test Plan which can be constructed before the actual testing phase begins.
  
  Output: Reliable code which conforms to the specification.

  Unit: Test of individual modules/objects/functions with white, sometimes black, box test.
    White Box: 
      Using the source code, develop test cases that will exercise the code in various
      ways: execute all paths, test boundaries of loops, etc.

    Black Box:
      Using the specification of the interface of the module, develop test cases that
      generate different inputs and check the output results.

  Integration:
    Test of modules until the software is completely integrated.
    May use:
      Top-Down : need stubs for underlying modules.
      Bottom-Up: need drivers to test underlying modules.
      Sandwich : spread out towards top and bottom with drivers and stubs.
      Builds   : groups of modules that can be tested together.
      Harness  : support environment (databases, etc.) to execute modules. 

  Validation:
    Test the software against the Validation Test Plan. 
    Does the software give the expected results using black box testing?

  System:
    Integrate and test the software with all system/hardware components.
  
  Performance:
    Measure the performance against specification, often using instrumentation.

  Stress:
    Run the system with abnormal quantity, frequency, or volume.

  Regression:
    Test a new release of software against a previous release and its outputs.
  
Maintenance:
  Goal  : Provide for the continued support of released software.
  Inputs: Bug Reports, Enhanchement Requests.
  Output: Updates and new versions of software.
  
    Rework:
      Correct, modify, enhance existing functionality.

    Reverse Engineering:
      Extract the design of the software from existing sources.

    Reengineering: 
      Modify the software to use more modern techniques.