22
From Defect Reporting to Defect Prevention A Lean Approach to Software Testing Sune Gynthersen & Lars Thorup BestBrains

From Defect Reporting To Defect Prevention

Embed Size (px)

DESCRIPTION

A talk given at DANSK-IT's Softwaretest 2010 conference.

Citation preview

Page 1: From Defect Reporting To Defect Prevention

From Defect Reporting to Defect Prevention

A Lean Approach to Software Testing

Sune Gynthersen & Lars ThorupBestBrains

Page 2: From Defect Reporting To Defect Prevention

Who are we?

Sune GynthersenLean/Agile software consultant

Lars ThorupFounder

Page 3: From Defect Reporting To Defect Prevention

Cost of correcting defects

Price

Time

Page 4: From Defect Reporting To Defect Prevention

A brilliant solution

Page 5: From Defect Reporting To Defect Prevention

No bashing of Waterfall today

"All models are wrong; Some models are useful"- George Box

Page 6: From Defect Reporting To Defect Prevention

A quick status...

• Number of currently open defects?o 10?o 50?o 100?o 1000?

Page 7: From Defect Reporting To Defect Prevention

A quick status...

• Prioritization of defects?o Priority 1o Priority 2o Priority 3

Page 8: From Defect Reporting To Defect Prevention

Is it really that bad?

"Current software projects spend about40 to 50 percent of their effort on avoidable rework"

- Barry Boehm (2001)

Page 9: From Defect Reporting To Defect Prevention

A true story (2010)

• 9000 defectso 3 hours per defect (reporting, prioritization, fixing, retest, accept)o 6 effective hours a dayo 220 work days a year

        = 20 man-years!

Page 10: From Defect Reporting To Defect Prevention

It reminds me of...

How many americans does it take to make a toast?

Page 11: From Defect Reporting To Defect Prevention

It reminds me of...

How many americans does it take to make a toast?

Two! One to burn it, one to scrape it

Page 12: From Defect Reporting To Defect Prevention

What is the cost of delay?

• A one month delay?• A one year delay?

Greater salary expenses

Lower financial return

Decreasing organisational inertia

Decreasing trust

Short-term prioritization

Page 13: From Defect Reporting To Defect Prevention

Vi believe...

• Any undesirable result represents a starting point for generating new learning.

Page 14: From Defect Reporting To Defect Prevention

We believe...

Or in plain english...

Every problem is an opportunity

Page 15: From Defect Reporting To Defect Prevention

The small details...

• Test versus Quality Assurance?

Page 16: From Defect Reporting To Defect Prevention

The small details...

• Testo Verificationo Test-drivingo Something that should

be automated

• Quality Assuranceo Early involvemento Proactivityo Help building the right

system in the right  quality

Page 17: From Defect Reporting To Defect Prevention

Conventional wisdom

Test/QA should find defects

Page 18: From Defect Reporting To Defect Prevention

A new hypothesis

Prevention is far more effective than fault-finding.

Rationale?

Page 19: From Defect Reporting To Defect Prevention

Focus

IfPrevention is far more effective than fault-finding

That meansQuality should be built-in from the beginning

Which should produceSignificantly higher profitability

Page 20: From Defect Reporting To Defect Prevention

What factors influence the quality?

Page 21: From Defect Reporting To Defect Prevention

What can I do?

Do more of this• Strive to create a common understanding

o no later than just before development starts• Exploratory testing

o as soon as possible• Stop feature work

o until defects have been corrected• Find, understand and remove the causes behind defects

 Do less of this• Manuel regression testing

Page 22: From Defect Reporting To Defect Prevention

4 questions for you...

• What is the most important output from Test/QA?• What is the avg. lead-time for defects? ("found" to "fixed")• How much undiscovered rework does your team have?• How do you reduce the amount of undiscovered rework?