Lecture 7: Software Quality

Lecture 7: Software Quality

• The lecture is about software quality, quality assurance, quality control, and reliability concepts.

Outline

• Software Quality
• Quality Assurance
• Quality Control
• Reliability Concepts

What is Software Quality Assurance?

• (No specific details given)

What is Quality?

• Quality is when a developed product meets its specifications.
• Problems can arise when:
  • Developer requirements exceed customer specifications, increasing development costs.
  • Certain quality characteristics (like maintainability) are difficult or impossible to define precisely.
  • Even if a product meets specifications, users may not consider it high quality if they weren't involved in defining the requirements.

Quality

• Quality is defined as "fitness for purpose."
• The customer decides if a product has quality.
• Easy-to-observe qualities include ease of interaction, graphical user interface (GUI) design, and freedom from errors.
• Harder-to-observe qualities include security and future maintainability.

Quality (cont.)

• Customers don't always know what they want.
• Asking customers what they do, instead of what they want, is sometimes more helpful.
• Customer desires regarding software quality are often not listed as top 10 features. Instead, important quality factors are reliability, usability, consistency, etc.
• Building quality software hinges on meeting customer needs, not just what they explicitly say they want.

What is Quality Management?

• Quality management ensures the required level of product quality is achieved.
• Procedures and standards are defined and applied in the product and process.
• Checking the follow-up of procedures is important.
• Collecting and analyzing quality data is crucial.
• Intangible aspects of software quality (like elegance and readability) are hard or impossible to standardize.

Software Quality Assurance (SQA)

• SQA is more than just testing—it's a collection of practices designed to minimize low-quality product risks.
  • Quality is "fitness for purpose"
  • Assurance is to be assured, be certain, and prevent defects.

What are SQA, SQP, SQC, and SQM?

• SQA is a systematic approach to ensuring software products and processes meet quality standards and user needs.
• SQP involves selecting and adapting appropriate procedures and standards based on specific software project goals & quality objectives.
• SQC defines and enforces processes to ensure quality procedures and standards are followed by the team.
• SQM collects and analyzes quality data to predict and control software product quality.

SQA Cont.

• SQA encompasses a variety of activities, including assessing products, managing testing, planning and development testing, automation, execution, QA project management, defect management, triage, metrics, and audits.
• It includes various processes like assessment, improvement, risk management, estimating and scheduling, release criteria, defect reporting, release process, project retrospection, change and peer review, verification and validation.

The Software Quality Iceberg

• There are internal and external aspects of software quality, which are visible and invisible.
• Internal quality aspects include coding practices, testability, maintainability, readability and flexibility; while external quality factors include correctness, efficiency, reliability, and program structure.

Why are Standards Important?

• Standards encapsulate best practices.
• They provide a framework for quality assurance processes.
• They ensure work continuity regardless of who is doing the work.
• While standards are important, they should be tailored to the specific project, if they are overly extensive.

SDS a Simplistic Approach

• ISO is not the only source of software development standards.
• Independent process and product standards are established.
• Product standards are not created by SQA.

Quality Standards

• (Specific certification details from document)

ISO - 9001 Elements

• Lists many elements for quality system requirements, software quality responsibilities, and more.

Process and Product Quality

• The quality of the development process directly impacts the quality of the final products.
• A factory-style approach doesn’t work for software; software is designed, not manufactured.

Obtaining Quality

• Quality control measures the quality of products produced (e.g., integration testing and code reviews).
• Quality assurance focuses on the process (e.g., ensuring all testing is conducted thoroughly).

Software Quality Control

• Methods for overseeing the software development process to ensure standards are met.

Methods of Software Quality Control

• Quality reviews—in-process assessments of processes and products to ensure quality. Quality reviews are responsible for writing Quality Engineering Records.
• Testing—assessments conducted after development, focused on verifying product results.
• Quality Audits—in-process process verifications to ensure compliance with standards.

Quality Reviews (cont.)

• Peer reviews involve groups of people reviewing processes and products. Thorough pre-review preparation is essential.
• Walkthroughs are reviews of products where participants are usually less formally prepared (e.g., tracing requirements to test procedures).
• Desk inspections are reviews by individuals where they submit feedback informally and individually.

Tests (cont.)

• Engineering dry-runs are performed by engineers and used to vet test procedures and the final product's readiness for release.
• SQE dry-runs are conducted by the software quality engineer (SQE) to verify the end product before formal customer testing.
• TFR or (run-for-record) tests are performed with the SQE and the customer to conclude testing the end product before sale.

Quality Audits (cont.)

• SQE audits are conducted by the SQE to ensure adherence to process standards (e.g., compliance, control, and management). QER forms document these results which are then escalated to upper management.
• Independent audits are conducted by independent entities to verify process standards (usually at a divisional or facility level).

Defect Detection

• Formal bug identification involves quality reviews and testing.

A Bug's Life

• Detailed process flows of how a bug is tracked and resolved.

Defect Prevention

• Defect prevention establishes practices to reduce reliance on bug detection, focusing on the broader process, metrics, risk-analysis, and control activities to prevent defects from occurring in the first place.

Continuity and Independence of SQA

• SQA teams need independence to provide objective feedback on processes.
• Processes may need to be tailored based on software specifics.
• Standards need to consistently apply regardless of task size or type.
• Quality is the responsibility of everyone.

Continuity and Independence of SQA (cont.)

• SQA teams should assist software engineers in their quality assurance and quality control activities.
• SQA teams do not directly work on projects.
• Their role is to help software engineers produce high-quality products..

The SQA Plan

• An SQA plan is created at the start of the project.
• Components include evals, reviews of standards, error tracking procedures and reports, and feedback to the project team

SQA Activities

• SQA supports selecting the appropriate process, reviews of software engineer activities & work products, ensuring correct deviation reporting, and documentation of any non-compliance. Feedback may be relayed to management.

What is Reliability?

• Reliability is a broad concept measured as a key aspect of quality, particularly related to system operation and user experience. If users rarely experience failures, the system is considered more reliable.
  • System reliability is a user-centric quality measurement.
  • Fault-free systems are the most reliable.
  • "Acceptable" failure rates are important aspects of reliability.

What is Reliability? (cont.)

• Failure occurs when the observed outcome of a program execution differs from the expected.
• A fault is the identified cause of a failure, often stemming from a defective code block.
• Time is a significant concept in evaluating reliability; short time gaps between failures result in lower reliability.
  • Execution time, or Calendar time, contribute to reliability evaluations.

Definitions of Software Reliability

• A first definition of software reliability is the probability of failure-free system operation over a specified time and environment. The definition contains key elements of probability, time duration, and operating environment.
• A second definition of software reliability is the failure intensity; a measure of reliability in a specific operation environment. Measured by frequency of failures over a specific period.

Factors Influencing Software Reliability

• A user’s perception of reliability depends on the number of faults and how users interact with the system.
• Factors influencing fault count include code size and complexity, development process characteristics, developer skills, and the operational environment.

Q & A, Any Questions?