1. Metrics "A science is as mature as its measurement tools."
-- Louis Pasteur

2. Starter Questions What can we measure?
What value can those numbers have?

3. Why Measure? quality improvement accurate estimation
what do we do well
what do we do poorly
accurate estimation
how productive are we

4. Types of Metrics Product Metrics Project and Process Metrics
direct measures - number of bugs, LOC
indirect measures - usability, maintainability
Project and Process Metrics
direct measures - costs, LOC per month
indirect measures - quality assurance, reliability

5. Code Metrics Size Efficiency Complexity Maintainability Lines of Code
Function Points
Efficiency
BigO
Complexity
Cyclomatic Complexity
Halstead's complexity metrics (next slide)
Maintainability

6. Halstead's Complexity Metrics
n1 = the number of distinct operators
n2 = the number of distinct operands
N1 = the total number of operators
N2 = the total number of operands
Program length N = N1 + N2
Program vocabulary n = n1 + n2
Volume V = N * (LOG2 n)
Difficulty D = (n1 / 2) * (N2 / n2)
Effort E = D * V

7. McCall's Quality Factors
Portability
Reusability
Interoperability
Maintainability
Flexibility
Testability
Product
Revision
Product
Transition
Product Operations
Correctness
Reliability
Efficiency
Integrity
Usability

8. Usability Integrity Efficiency Correctness Reliability Maintainability
Operability
Training
Communicativeness
Input/Output volume
Input/Output gate
Access Control
Access Audit
Storage efficiency
Execution Efficiency
Traceability
Completeness
Accuracy
Error Tolerance
Consistency
Simplicity
Conciseness
Instrumentation
Expandability
Generality
Self-Descriptiveness
Modularity
Machine Independence
Software System Independence
Communications Commonality
Data Commonality
Usability
Integrity
Efficiency
Correctness
Reliability
Maintainability
Testability
Flexibility
Reusability
Portability
Interoperability

9. ISO 9126 Quality Characteristics and Guidelines for Their Use
Quality Factors
Functionality
Reliability
Usability
Efficiency
Maintainability
Portability

10. Design Metrics Fan In Fan Out Morphology
based on number of nodes, depth, width

11. Module Design Metrics Cohesion Coupling
how many functions does a module perform
coincidental
logical - eg does all output
temporal - eg all the startup work
procedural - executed in this order
communicational - module arrangement on work on what data
functional
Coupling
how is the module connected to other modules
global variable, parameters, stands alone

12. Object-Oriented Metrics
Weighted Methods per Class
not only how many methods per class are there, but also how complex are they
Depth of Inheritance Tree
Number of Children
how many child classes does a class have
Response for Class
number of local methods, plus number of methods they call
Lack of Cohesion Metric
number of non-intersecting (don't use the same variables) methods

13. Lack of Cohesion Example
Module A calls Module B
B accesses Variable X
C and D access Y
D calls E
This should be split into two classes. A C D B E Y X

14. Project Metrics LOC or FP per month
Errors per LOC (aka Defect Density)
Defect Removal Efficiency
Time required to make changes
Test coverage
Required Skills

15. Obviously, Beware of Statistics
Version Version 2
Module# Defects LOC Defects/LOC Defects LOC Defects/LOC Relation
<
<
<
<
<
Sum >

16. Basic Questions
What are the basic metrics that managers need to track?
How do we gather all these numbers?
When do we process all these numbers?

17. SEI CMM Level 2 (repeatable)
Software Requirements Management
status of allocated requirements
number of changes to requirements
Software Project Planning
completion of milestones compared to the plan
work completed, funds expended, … compared to plan
Software Project Tracking and Oversight
resources expended to conduct oversight

18. SEI CMM Level 3 (defined)
Training Program
number of training waivers approved
actual attendance vs projected attendance
results on post-training tests
Software Product Engineering
numbers, types, and severity of defects by stage
effort to analyze proposed changes
number of changes by category

19. Summary
To do something well, we must understand what we are doing. To understand something, we must be able to measure it.
We can measure what we are building and we can measure our building process.

20. Next Time… Testing "Homework" Assignment:
find a testing tool, any testing tool
how much does it costs
what does it do …