Chapter 10: Quality of Deliverables and Products

Question 1

C10 Quality of Deliverables and Products: Terms and terminology

Answer 1

• Agile особенности - постоянно мониторится quality в каждой итерации и на каждом митинге. Чем позже выясним отхождение, тем хуже
• Audit / Inspection for Quality - on the exam “Audit” related to Manage Quality, “Inspect” - Control Quality. We audit processes and we inspect product or deliverables
• Defect Cycle Time (agile/hybrid) - время от обнаружения до устранения
• Grade/Quality
  
  • Grade - какой вес выдерживает выбранный нами вид бетона
  • Quality - какого качества должен быть бетон
• Gold plating is the process of adding extra features to drive up costs and consume the budget.
• Normal distribution - expressed as a chart, the shape of a bell curve. Used to measure variations away from the “norm”
• Mutual exclusivity - two events - if they cannot both occur in a single trial (e.g. you cannot see at the same time both sides of the coin)
• Quality - degree to which a project fulfill requirements (!)
• Quality Department - assume that the company has one. Помогает во всех процессах
• Statistical independence - probability of one event doest not affect the probability of another event
• Standart deviation (or Sigma) - metric for a range of measurement. Shows how far a measurement is from the average.

Question 2

Differences between Plan, Manage and Control Quality

Answer 2

• Manage - focus on processes, procedures and standards
• Control - quality of deliverables

Детали - стр.254

Question 3

Methods for Planning Quality / Managing Quality / Controlling Quality

Answer 3

Methods for Planning Quality

• Interviews, Brainstorming, Benchmarking
• Decision-making Methods
  
  • Multicriteria decision anaylsis
  • Prioritization diagram
• Cost-benefit Analysis
• Cost of Quality (COQ) - отдельная карточка
• Marginal Analysis - finding the point at which the benefit from improving quality quals the cost to achieve it. When that point is reached, PM stops trying to improve quality
• Logical Data Models - represent types of data in a particular area, and the relationships between those data types (два связанных квадрата, в каждом набор данных/факторов)
• Meetings
• Mind mapping - used to facilitate gathering of quality req and illustrate their impacts on other parts of proj planning
• Matrix Representations - например list of quality req in one column, and their characteristics (e.g. priority) in others
• Prioritization Matrix (Chart) - effort/impact matrix to decide about quality planning
• Flowcharts - used to: define processess, show dependencies in a process, study the steps of a process that is causing a quality defect
• Test and Inspection Planning
  
  • Determining how the team will confirm that the required level of quality has been achieved in the completed proj deliverables
  • How the deliverables will be evaluated for perf and reliability

………………………………………..

Methods for Managing Quality
To analyze the processes

• Checklists - steps of the process completed
• Cause-and-Effect (Fishbone, Ishikawa, or Why-Why) Diagrams - confirm that policies and procedures are being followed and metrics are being used correctly
• Scatter Diagrams
• Histograms - help to identify frequency of defects
• Document analysis- reviewing the results of testing and other quality reports
• Alternatives analysis - e.g. - would a new automated testing tool be more beneficial?
• Design of Experiments (DOE) - An approach that relies on statistical scenarios to determine what variables have an optimap impact on the proj deliverables. Пробуем, тестируем, потом распространяем
• Flowcharts
• Process Analysis - ищем способы improvement. Часто в repeatable processes
• Root Cause Analysis
• Failure Analysis
• Multicriteria Decision Analysis
• Affinity Diagrams - helps to group and organize the results of root cause analysis
• Audits
• Design for X (DfX) - ищем что изменить чтоб улучшить качество конкретного аттрибута quality
• Problem-solving
  
  1. Define the real rootcause
  2. Analyze
  3. Identify solutions
  4. Pick a solution
  5. Implement a solution
  6. Review the solution and confirm solving the problem

………………………………………..
Methods for Controlling Quality

Inspect and verify (not audit!) that deliverable meets req stated in Quality Management Plan, and can be moved to Validate Scope (end user acceptance)

• Checklists - проверяем что все features включены. При этом сам лист мог готовиться в Manage Quality
• Root cause analysis
• Cause-and-effect diagrams
• Scatter diagrams
• Histograms and Pareto charts - identify problems that need the most immediate attention

Отдельно только для Control Quality:

• Checksheets
• Meetings
• Statistical Sampling - берем сет 20 единиц из 100 и на основе кол-ва фейлов понимаем, какой процент их в общем числе и нужно ли смотреть остальные
• Questionnaires and Surveys
• Project Performance Reviews comparing the results of control measurements to metrics in Quality Management Plan
• Inspection - walk-throughs, measurement of deliverables. В процессе могут использовать Checklists/Checksheets
• Control Charts - отдельная карточка

Question 4

Cost of Quality (COQ)

Answer 4

Измеряем, что не тратим слишком много, чтобы достичь определенного уровня quality.

Categories of costs associated with quality:

• Prevention - costs of preventing any quality issues from occurring
• Appraisal - costs for monitoring and controlling (e.g. quality audits etc.)
• Internal failure - нашли ошибки before the product reaches the customer (e.g. waste, scrap, rework)
• External failure - нашли ошибки после передачи клиенту. Impact goes beyond money to reputation

Оно же на картинке Cost of Conformance / Cost of Non-conformance

Question 5

Control Chart

Answer 5

Знать для экзамена термины
Используется для quality, cost performance, schedule variances

Пример: The drug must stay effective in the system for a minimum of three hours but last no more than five hours, to prevent accidental overdose.

• Plotting the control chart - вычерчивание кривой по точкам. Берем samples, затем The chart shows whether each sample is within acceptable limits
• Upper and lower control limits (“tolerances” in agile) - определяют, whether data is IN control or OUT OF control
• Mean (average) - центральная линия. При этом Normal Distribution Curve (парабола) показывает acceptable range of variance around a mean
• Upper and lower specification limits - устанавливаются клиентом, ожидания клиента/контракта. Должны быть ВНЕ control limits
• Out of control (“out of tolerance” in agile)
  
  1. Вылетели за control limits
  2. Non-random data points вылезают на одной стороне - см. Rule of Seven
• Rule of Seven (heuristic rule) - 7 подряд значений на одной из сторон от MEAN (Average). Означает что процесс out of control
• Assignable cause / special cause variation - показывает что process out of control, если точка вылетела за control limit. Нужно исследовать и исправлять

Question 6

Work in progress (WIP) and Cycle time

Answer 6

• Lead time - length of an entire process
• Cycle time
  
  • length of a part of the process (see pic)
  • = WIP / Throughput (среднее или velocity per iterations, не до конца понял)
  • если много - значит слишком большой WIP. Это плохо

Question 7

Quality metrics

Answer 7

Чтоб понимать, как хорошо реализуется проект и когда делать changes, на этапе Plan Quality Management надо определить метрики, и приемлемый range of variation. Примеры:

• Number of changes
• Variance related to resources utilization
• Number of items that fail inspection
• Variance of the weight of a product produced by the project compared to the planned weight
• Number of bugs found in software being developed