Quality is Job 1

In the Beginning…

• Deming – the name of quality

• Think of autos

• In software– Program mistakes– Development time– Originally inspect final product; identify

errors; and then work to repair damage

In a Modern System

• Now…– Design a quality system up front– Build quality into the product– Filter out errors as early in the project as

soon as detected– Also – define the level of quality that is

desired

A Balance…

• Must be reached between quality and resources

• Processes should be continually improved along the way

In PMBOK

• Quality Planning

• Perform Quality Assurance (user requirements)

• Perform Quality Control (meets standards and eliminate unsatisfactory performance)

10 Principles for Quality in Software

Following the principles will generally ensure quality in your software

Principle 1-Use Feedback

• Defect Prevention Process (DPP) = Software Engineering Institute's Capability Maturity Model (CMM) Level 5

• Inspection Method – IBM

• Evolutionary Project Management

• Statistical Control

Principle 2 – Identify Critical Measures

• For each critical factor you would find a series of faults that would include the following: failure to..– Systematically identify all critical

stakeholders and their critical needs– Define the factor measurable– Define a practical way to measure the factor– Contract measurably for the critical factor

Cont…

• Design toward reaching the factor’s critical levels

• Make the entire project ream aware of the numeric levels needed for the critical factors

• Maintain critical levels of performance during peak loads or on system growth

• As a matter of fact most quality factors are not defined quantitatively at all (only performance, budget and deadlines are) (Others that are not defined are serviceability, reliability, adaptability, portability and usability)

Principle 3 – Control Multiple Objectives

• Must be able to deal with all critical factors simultaneously

• Use Impact Estimation Table – tracking estimation against goals

• When deviations are noted then you can control immediately for them

Impact Estimation TableStep 1

Plan 1

Step 1

Actual

Step 1

Difference

Step 1

Total

Step 2 and so forth ….

Reliability 50%

+_ 50%40% -10% 40%

Continue on with totals for cumulative stepsPerformance

11 sec-1 sec

80%

+- 40%40% -40% 40%

Usability

30 min-30 sec10%-

+_20%12% +2% 12%

Capital Cost

1 million20% -+_10%

10% +10% 10%

Programming Hrs 10,000

2%-+-1%

4% -2% 4%

Calendar Time 1 wk 2 wk -1 wk 1 wk

Principle 4 – Evolve in Small Steps

• Dealing with the unknown and risk

• Goal is to deal with development in small steps

• Not necessarily modules, but those items that satisfy stakeholders

• May be as small as a day or week

Principle 5 – A Stitch in Time Saves Nine

• Need strong specification standards

• Keep reducing minimum until acceptable

• Use inspection sampling to reduce errors

Principle 6 – MotivationMoves Mountains

• Four categories of problems– Will to change– Knowledge to change direction – channel energies

• Measurable, quantified clarity of the requirements and objectives of various stakeholders (people generally unclear about goals and no ability to get numeric feedback)

• Knowledge of all the multiple critical goals (focus too much on one goal and not on all)

• Formal awareness of constraints such as resources and laws– Ability to change– Feedback about progress in the desired change

direction

Principle 7 – Competitionis Eternal

“We have an infinite set of milestones or evolutionary steps of result delivery and use them as we need; the moment we abandon a project, we hand opportunity to our competitors. They can sail past our levels of performance and take our markets.”

Principle 8 – Things Take Time

• “It takes two to three years to change a project, and a generation to change a culture.”

• There needs to be a long-term plan with visible short-term progress.

Principle 9 – The Bad with the Good

You must have– A quantified, multidimensional specification of our

requirements, our quality objectives, and our resources (people, time, $$)

– Knowledge of the expected impact of each design idea on all these quality objectives and resources

– Evaluate each design idea with respect to its total – expected or real – impact on requirements, unmet objectives and the unused cost budgets

Principle 10 – Keep Your Eyes on Where You Are Going

• To discover the real problem – ask of a specification – Why?

• Task is to design and deliver high quality software that meets the specifications!

• Keep people motivated!