27
Ph. D. candidate Ph. D. candidate Scuola Superiore S.Anna Scuola Superiore S.Anna Pisa - Pisa - Italy Italy Designing Real-Time Designing Real-Time Software Software visiting prof. Baruah visiting prof. Baruah Univ. of North Carolina Univ. of North Carolina Chapel Hill - NC Chapel Hill - NC

Software Quality Assurance II Due today: Detailed Design Document I Next Class:Pressman 20; Quiz #2 Questions? / Team Status Reports Continuous Improvement

  • View
    212

  • Download
    0

Embed Size (px)

Citation preview

Software Quality Assurance II Due today: Detailed Design Document I

Next Class:Pressman 20; Quiz #2

• Questions? / Team Status Reports

• Continuous Improvement (cont.)

• SQA in the real world

• Bio Break (5 minutes)

• More SQA in the Real World

• Quiz #2 Review

Success Factors for Effective Continuous Improvement

• Feedback & Communication– Assessment Results

• Take Action on most important issues– Prioritize issues– Formulate action plans– Execute actions

• Re-assess results of actions

Assessment of this Class

On a blank piece of paper, please do the following (anonymously):

1. Describe at least one area where Craig is doing well?

2. Describe at least one area where Craig could improve ?

3. Rate Craig’s overall performance on a 1-5 scale (1=poor, 5= excellent).

4. Rate your overall satisfaction with the course on a 1-5 scale (1=very dissatisfied, 5= very satisfied).

Results: Numerical Ratings

• Overall Rating Craig: 4.4– Standard deviation: 0.48

• Overall Rating Course: 4.6– Standard deviation: 0.75

Results: Craig – Strengths & AFIs

Strengths (n=30)• Real world examples/applications (40%)

• Good / clear lectures (33%)

• Involves class (13%)

• Good overheads (10%)

• Highlights book sections in lecture (3%)

AFIs (n=18)• Clarify / highlight quiz material (39%)• Post detailed lecture notes on web (11%)

• Change homework each week (6%)• More graphics in slides (6%)• More real world examples (6%)• Add guest lectures (6%)• Slow down class (6%)• Keep class moving faster (6%)• More time on architecture (6%)• Get more involved with projects (6%)• Unrealistic schedule for 1st demo (6%)

Action Plan

More SQA in the Real World

• #1 Motivation for SQA Efforts

• #5 IBM’s “Train Process”

• #6 Rigorous Code & Design Reviews

• #8 Defect Prevention Process (DPP)

Secrets of Software Quality

25JAIIOCraig Kaplan, Ph.D.

I.Q. Company

http:\\www.iqco.com

1

20% Savings on Service Costs

0%2%4%6%8%

10%12%14%16%18%20%

% Savings in Service

Costs

1990 1991 1992 1993

Cost Savings at IBMSTL

Source: Secrets of Software Quality by Kaplan, Clark, & Tang (McGraw-Hill 1995)

56% Increase in Revenue per Employee

0%

20%

40%

60%

80%

100%

120%

140%

160%

% 1989 Baseline

1989 1990 1991 1992

Revenue per Employee atIBM STL

Source: Secrets of Software Quality by Kaplan, Clark, & Tang (McGraw-Hill 1995)

14% Improvement in Customer Satisfaction

85%

90%

95%

100%

105%

110%

115%

% of 1991 Baseline

1991 1992 1993

IBM STLCompetitors

Source: Secrets of Software Quality by Kaplan, Clark, & Tang (McGraw-Hill 1995)

46% Reduction in Field Defects

0%10%20%30%40%50%60%70%80%90%

100%

% of 1989 Baseline

1989

1990

1991

1992

1993

Defects at IBMSTL

Source: Secrets of Software Quality by Kaplan, Clark, & Tang (McGraw-Hill 1995)

5

Collapsing the Development ProcessBusiness Case

Initial Specs.

Final Specs.

Coding

Ext. View Doc.

Test Case Dev.

Formal Test

Business Case

Initial Specs.

Final Specs.

Coding

Formal Test

Ext. View Doc.

Test Case Dev.

Reduced Cycle Time

The “Train” ProcessRelease 2.0

Ships on May 1.Feature #1 for Rel. 2.0

Feature #1 Team: Programmer, Tester, Technical Writer

6

Typical Defect Depletion Curve

0%

5%

10%

15%

20%

25%

30%

0 I II III IV V VI VII

Defects

Defects

Found

Design Code Unit Formal Ship +6 +18 Test Test +3 Mo. Mo. Mo.

60% 40%

Shifting the Defect Depletion Curve

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

0 I II III IV V VI VII

Defects

Defects

Found

Design Code Unit Formal Ship +6 +18 Test Test +3 Mo. Mo. Mo.

90% 10%

Reviews & Inspections

Results from IBM Study

Code Type: Ship Tested New TotalInspection Sessions 11 10 40 61Major Errors 22 59 376 457Minor Errors 51 123 1267 1441Total KLOC 4 7 36 47Total Hours 197 386 1017 1600Maj. Errors/KLOC 5.5 8.4 10.4 9.7Hours/Maj Defect 9.0 6.5 2.7 3.5

Hours/Maj Defect spent in Formal Test: 15 - 25

8

How DPP Works

Centralized data collection and storage tools

Ongoing Work

Causal Analysis Team

Action Team

Kickoff Meetings

defects

suggested actions

improvements

apply improvements, add feedback

Some DPP Success Factors

• Need for DPP clear

• Management commitment

• DPP education for managers & developers

• Prioritization by Action Team

• Rotate Action Team Assignments

• Administrative support for defect logging

• Continuous improvement of DPP

Quiz #2 Format / Coverage

You are responsible for all of the lectures and reading to date.

• Format will be the same as Quiz #1:– Multiple choice

– Short answer

– Essay

• 1/3 covering material that Quiz #1 covered

• 2/3 covering new material since Quiz #1

Quiz #2 Review Topics

• All topics of Quiz #1 review sheet• WebE Principles, questions, issues• Web UI guidelines• Software design model• Stepwise refinement• Modularity, cohesion, coupling• UI – 3 golden rules• Task analysis• Design issues / principles• Data warehouses• Data design at the component level• Basic Architecture styles/types

• Cost of quality• SPC• Defect Prevention Process• Inspection and reviews• SQA metrics• Quality standards• Continuous Improvement• SQA real-world applications• Change Control process• All questions on Quiz #1