Software Engineering Qualities - Computer Science, Engineering Design - Lecture Slides, Slides of Computer Science

Download Software Engineering Qualities - Computer Science, Engineering Design - Lecture Slides and more Slides Computer Science in PDF only on Docsity!

Chapter 2: SWE Qualities

Software Engineering Qualities

• Software is a Malleable Product that is Human

Intensive (High People to Machine Ratio)

• Software Engineering Qualities
  • Correctness and Reliability
  • Robustness and Performance
  • User Friendliness and Verifiability
  • Maintenance and Reusability
  • Repairability and Evolvability
  • Portability, Understandability, and Interoperability
  • Productivity, Timeliness, and Visibility
• Requirements for Information, Real-Time, Distributed,

and Embedded Systems

The Correctness Software Quality

• Functional Correctness: Software Behaves According to Requirements Specification
• Assessing Correctness - Potential Problems
  • Assumes Requirements Specifications are Available and Up-to-Date
  • Assumes Correctness Testing is Possible
  • Assumes Mathematical Properties can be Established Between Software and Specs
• Correctness - Absolute Quality (Yes/No)
• Correctness in Practice
  • Focus on Components at Varying Levels
  • Addressed via Public Interface, Private Implementation, Testing of Classes/Components
  • Components in Distributed Apps

The Reliability Software Quality

• Software Performs as User Expects
• User Can Depend on Software
• Different Reliability Requirements per Domain
  • Computer Games: Low?
  • Flight Control Software (Mars): High?
  • Banking, Insurance, Stocks, …: High?
• Correct Software is Reliable but Reliable Software May Not be Correct, that is …
• Reliability in Practice:
  • Focus on Reliability of Components (Classes) and their Composition into Systems
  • Critical for Commercial Success of Consumer Products and Services

Correctness

Reliability

The Performance Software Quality

• Performance is Equated with Efficiency, Scalability,

Reliability, User-Acceptance, etc.

• Growing Problem Impacts Strongly on Software
  • Varied OS Platforms (NT, Win2K/XP, Solaris)
  • Different HW Capabilities (Memory Size, 32 vs. 64 bit, Disk Capacity, Display, etc.)
• Measurement vs. Queuing Analysis vs. Simulation
• Impact from Specification through Delivery
• Performance in Practice
  • Identify Key Classes/Components with Hard Performance Constraints - Embedded Applications (Elevators, Avionics, etc.)
  • Versions of Classes/Components for OS/HW Platforms?

The Performance Software Quality

• Measurement: Measure Actual Performance of Working

Systems

• Analysis: Build a Model and Analyze
  • General Understanding of System
  • Queueing Models During Specification/Design
  • Redo Queueing Models After Implementation by Providing Hard Numbers to Fine-Tune
• Simulation: Model Simulates Actual System
  • Predict Exact Performance of Components
  • Expensive and Time Consuming to Build
  • Yields an Accurate Estimate
• System Performance Must Be Addressed from Day One

Throughout Lifetime

Maintenance of Web Applications

• Three Types of Maintenance
  • Corrective, Adaptive, and Perfective
  • Need to have Strategy for Making Changes
• Web Apps Typically Developed in Staged Setting
  • Development/Testing Environment Web + Application Server on Intranet
  • Actual Web + Application Servers on Internet
  • May be Replicated Web/Application Pairs that are Geographically Separated
• Changes/Testing on Intranet
• Migration to Internet in a Staged Manner for Upgrades
• Note:
  • Maintenance Requires Version Management
  • Source Code Control System

The Reusability Software Quality

• Strive for Domain-and-Organization Specific Reuse
• Products Most Likely to be Developed in Future
• Focused on Software Components
  • Scientific Math and String Libraries
  • Windowing and GUI Building Blocks
  • Java APIs for Enterprise, Embedded, etc.
• Reusability Evolving to Larger Components (e.g., Java Beans) and Subsystems
• Reusability in Practice
  • In OO, Achieved via Inheritance and Generics
  • Defined/Controlled by Designers and Developers
  • Components Play a Major Role in Reuse and in Distributed/Web Applications - Reuse “Shopping Cart, Checkout Procedures, etc.”

The Evolvability Software Quality

• Field of Dreams: If you Build it, it Will Change!
• Key Evolvability Issues
  • Ease of Modifications - Add New Features and Change Existing Features
  • Function of System Size
  • Modularization vs. Monolithic
• Over Time, Evolvability Reduces as Subsequent Versions of Same Software are Released
• Reach of Limit of Change w.r.t. Look-and-Feel
• Evolvability in Practice
  • Achieved via Inheritance and Polymorphism
  • Public Interface Must Remain Unchanged!
  • Adding Capabilities and Features to Web and Distributed Applications the Norm

The Portability Software Quality

• Software Works on Different Platforms
• Origins and Popularity of C and Unix
• Hard in Practice - All C++’s Not Created Equal
  • Borland C++ vs. Visual C++ vs. Sun C++
  • Java – Compiler Addons and Competing Versions
  • 32 vs. 64 bits
  • Core Language Same? (Operator Precedence)
  • Compiler/Product Specific Libraries
• Newer Languages (Ada95, Java) Stress Ability to Port without Rewriting Any Code
• Portability in Practice
  • Isolate Specifics into Dedicated Classes
  • Portability Transcends D & D Paradigm
  • Web – App works in Multiple Browsers/Versions

The Interoperability Software Quality

• Enterprise Computing: The Problem of Today and Future!
• New Products Must Coexist and Cooperate with Legacy,

COTS, Databases, etc.

• Success in Interoperability Traced to …
  • Programming Interface for COTS/Legacy
  • DCOM/OLE, CORBA, DCE, JINI, .NET
  • Emerging Languages: Java, Active-X, C#
• Open Systems, Standards, Multiple Vendors, etc.
• Interoperability in Practice
  • Focus on Public Interface Concept
  • Interoperability Transcends D & D Paradigm
  • Norm for Web and Distributed Applications

Typical Process Qualities

• Process Qualities Involve the Steps and Actions that are

Integral to Software Development

• Productivity
  • Denotes the Efficiency and Performance of Software
  • A Measurable Quality (Empirical)
• Timeliness
  • Ability to Deliver a Product on Time
  • Meeting Deadlines and High Quality Products
• Visibility
  • All of Its Steps and Current Status are Documented Clearly
  • Open Software Process (All Stakeholders)

The Timeliness Software Quality

• Meeting Deadlines and Market Demands
• Requires Careful Scheduling, Accurate Estimation, and Clear Milestones
• Impacted by
  • Changing User Requirements
  • Skills (or Lack Thereof) of SW Engineers
• Incremental Delivery
  • Product Released in Stages
  • Complete and Incremental Requirements
• Timeliness in Practice
  • Promotes Increments and Components
  • Classes Facilitate Evolution, Testing and Integration
  • For All Applications (Centralized to Distributed)

Timeliness: Visual Description of

Mismatch

• Often the Development Process Does Not

Follow the Evolution of User Requirements

• A Mismatch Occurs Between User

Requirements And Status of the Product Function

t 0 t 1 t^2 t 3 t 4

User needs Actual system capabilities

Time