1. Chapter 4 Software Process and Project Metrics

2. Measure
A measure is a mapping from a set of entities and attributes in the real word to a representation or model in the mathematical world.
One can manipulate the numbers or symbols in the mathematical world to obtain more information and understanding about the real world.

3. Measurement process process metrics project metrics measurement
What do we
use as a
basis?
• size?
• function?
project metrics
process metrics
process
product
product metrics

4. Measurement Advantage For software process For software project
For software engineer

5. Measures, Metrics & Indicators

6. Indicator Private indicator vs. Public indicator
Process and Project Indicator
assess the quality of an ongoing project
modify the technical approach to improve quality
adjust work flow or tasks to avoid the delay
track potential risks
uncover problem areas
evaluate the project team’s ability

7. Process Metrics
Provide indicators – lead to long-term software process improvement
Organization – gain insight into the efficacy of an existing process
Managers and Practitioners -- assess what works and what doesn’t

8. Statistical Software Process Improvement (SSPI)
Failure analysis
Categorize the defects
Record the correcting cost
Count and rank the defects
Compute overall cost in each category
Uncover the highest cost categories
Develop the plan to modify the process

9. Statistical Software Process Improvement (SSPI)

10. Fishbone Diagram

11. Process Metrics Guidelines
Grady
Use common sense and organizational sensitivity when interpreting metrics data.
Provide regular feedback to the individuals and teams who have worked to collect measures and metrics.
Don’t use metrics to appraise individuals.
Work with practitioners and teams to set clear goals and metrics that will be used to achieve them.
Never use metrics to threaten individuals or teams.
Metrics data that indicate a problem area should not be considered “negative.” These data are merely an indicator for process improvement.
Don’t obsess on a single metric to the exclusion of other important metrics.

12. Project Metrics Project Indicator Measure
Input -- sources required to do the work
Output -- work products created during the process
Results -- effectiveness of the deliverables
Project Indicator
Assess the Quality of an Ongoing Project
Modify the Technical Approach to Improve Quality
Adjust Work Flow or Tasks to Avoid the Delay
Track Potential Risks
Uncover Problem Areas
Evaluate the project team’s ability
Project metrics
Errors uncovered per review hour
Scheduled vs. actual milestone dates ….

13. Software Measurement
Direct measures: Defects, LOC produced, execution speed, memory size…
Indirect measures: quality, complexity, efficiency…

14. Typical Size-Oriented Metrics
Normalization value
Metrics:
errors per KLOC
defects per KLOC
$ per KLOC
page of documentation per KLOC
errors per person-month
LOC per person-month
$ per page of documentation

15. Typical Size-Oriented Metrics
Advantages
easily counted
Disadvantages
programming language
well-design
late get data

16. Function-Oriented Metrics
Function Point
countable (direct) measures of software’s information domain
assessments of software complexity

17. Function Points (FP) Establish count for input domain and
Analyze information
Establish
count
for input domain and
domain of the
application
system interfaces
and develop counts
Weight each count by
Assign level of complexity or
weight
assessing complexity
to each count
Assess influence of
Grade significance of external factors, F i
global factors that affect
such as reuse, concurrency, OS, ...
the application
function points =C x (count x weight)
Compute
function points
where:
complexity multiplier: C = ( x N)
degree of influence: N = F i

18. Function Points (FP)

19. Function Points (FP) Complexity adjustment values (Fi ; i = 1-14)
Does the system require reliable backup and recovery?
Are data communications required?
Are there distributed processing functions? :
14.

20. Function Points (Example)

22. Function-Oriented Metrics
errors per FP
defects per FP
$ per FP
pages of documentation per FP
FP per person-month )

23. Function-Oriented Metrics
Advantages
programming language independent
get data early
Disadvantages
subjective
difficult to collect the data
no direct physical meaning

24. Quality Metrics
There are different answers for the characteristics of software that contribute to its overall quality.
Users – external view
Practitioners – internal view
Build models to relate the user’s external view to the developer’s internal view of the software.

25. External quality
Product quality

26. Metrics for Software Quality
Correctness — operate correctly
(defects per KLOC)
Maintainability — amenable to change
(mean-time-to-change (MTTC))
Integrity — impervious to outside attack
(integrity = Sum( 1- threat *(1-security)))
Usability — easy to use
(learning time, productivity, assessment
of user…)

27. Defect Removal Efficiency (DRE)
For Project
DRE = (errors) / (errors + defects)
where
errors = problems found before release
defects = problems found after release
DRE = (errors) / (errors + defects)
For Process
DRE i = (errors)i / (errors i + errors i+1)

28. Metrics Baseline Attributes of baseline data reasonably accurate
as many projects as possible
consistent
similar

29. Managing Variation (Control Chart Approach)1 2 3 4 5 6 7 9 11 13 15 17 19
Project s E r , o f u n d / e v i w h
FIGURE 4.8 Metrics data for errors uncovered per review hour

30. Moving Range Control Chart (stable)
UCL mR
Mean of the moving Range (mR) = 1.71
Upper Control Limit (UCL) = mR * = 5.57

31. Individual Control Chart (control )
UNPL Am
LNPL
Upper Natural Process Limit (UNPL) = mR * Am = 8.55
Lower Natural Process Limit (LNPL) = mR * 2.66 – Am = 0.55
Standard deviation = (UNPL – Am) / 3 = 1.52

32. Individual Control Chart (control )
Out of control If any one of the following items is true
A single metrics value lies outside the UNPL.
Two out of three successive metrics values lie more than two standard deviations away from Am.
Four out of five successive metrics values lie more than one standard deviation away from Am.
Eight consecutive metrics values lie on one side of Am.

33. Metrics for Small Organizations
Keep the measurements simple.
Tailor measurements to each organization.
Ensure the measurements produces valuable information.
The cost of collecting and computing metrics ranges from 3% to 8% of project budget during the learning face, then drops to less than 1%.

34. Software Metrics What is it?
A quantitative measure of the degree which a system, component, or process possesses an attribute.
Why is it important?
Objective evaluation, better estimates, true improvement…
How does it?
Collected  analyzedcomparedassessed
Who does it?
Collect Measures  software engineers
Analyze and assess Metrics  software manager

35. Software Process and Project Metrics
What is it?
Who does it?
Why is it important?
What are the steps?
What is the work product?
How to ensure it be done right?