1. A Tale of Two XP Teams Laurie Williams North Carolina State University
“It was the best of times, it was the worst of times.”
Laurie Williams North Carolina State University
© Laurie Williams 2001

2. Agenda IBM Team: “Safe Subset” of XP Practices
Sabre Team: “Mostly all” of XP
Summary/Comparison

3. Extreme Programming Examination
Extreme Programming Evaluation Framework XP-EF (said X-pef)
XP-Context Factors (XP-cf)
XP-Adherence Metrics (XP-am) (said X-pam)
XP-Outcome Measures (XP-om) (said X-pom)
Reusable framework for reporting:
the extent to which an organization has adopted XP practices; and
the result of this adoption

4. IBM: XP-Context Factors (XP-cf)
Small team (7-10)
Co-located
Web development (toolkit)
Supplier and customer distributed (US and overseas)
Examined one release “old” (low XP) to the next “new” (more XP)

5. IBM: XP-Adherence Metrics (XP-am)
Subjective: Shodan Survey (
Old 56%
New 72%
Objective Metrics
XP-am Metric
Practice
Old
New
Automated test class per user story
Testing
0.11
0.45
Test coverage (statement)
30%
46%
Unit test runs per person day
14%
11%
Test LOC/Srce LOC
0.26
0.42
Accept test execute
Manual
Did customers run your acceptance tests? No
Pairing Frequency
Pair Pro
<5%
48%
Release Length
Shrt Rel
10 months
5 months
Iteration Length
Weekly

6. IBM: XP-Results Metrics (XP-rm)
XP Result Metric
Old
New
Internal Code Structure (mean values)
Methods per class
Depth of inheritance tree
Number of children
Coupling
Response for class
Lines of code per class
McCabe Complexity
Response to Customer Change (Ratio (user stories in + out) /total) NA
0.23
Internally-Visible Quality
(test defects/KLOEC of code)
1.0
0.50
Externally-Visible Quality
(released defects/KLOEC of code)
0.24
Productivity (stories / PM)
Relative KLOEC / PM
1.34
1.7
Customer Satisfaction
High
Morale (via survey)
1.11

7. Sabre: XP-Context Factors (XP-cf)
Small team (6-10)
Co-located
Scriptable GUI environment
Customer remote, multinational, several time zones
Examined third release “old” (low XP) to the ninth release “new” (sustained XP)

8. Sabre: XP-Adherence Metrics (XP-am)
Subjective: Shodan Survey
New 77%
Objective Metrics
XP-am Metric
Practice
Old
New
Automated test class per user story
Testing
N/A
3.22
Test coverage (statement)
32.9%
Unit test runs per person day
None
1.0 (anecdotal)
Test LOC/Srce LOC
0.054
0.296
Accept test execute
Manual
Did customers run your acceptance tests? No
Pairing Frequency
Pair Pro 0%
50%
Release Length
Shrt Rel
18 months
3.5 months
Iteration Length
10 days

9. Sabre: XP-Results Metrics (XP-rm)
XP Result Metric
Old
New
Internal Code Structure (mean values)
Methods per class
Depth of inheritance tree
Number of children
Coupling
Response for class
Lines of code per class
McCabe Complexity
Response to Customer Change (Ratio (user stories in + out) /total) NA
N/A
Internally-Visible Quality
(test defects/KLOEC of code)
1.0
0.25
Externally-Visible Quality
(released defects/KLOEC of code)
0.70
Productivity (stories / PM)
Relative KLOEC / PM
1.46
Customer Satisfaction
High
Morale (via survey)
68.1%

10. Summary Two characteristically-agile teams:
When used by teams operating within the specified context, the use of a specified subset of XP practices leads to an improvement in . . .
Alternative Hypothesis
IBM Case study evidence?
Sabre case study evidence?
internal code structure No
pre-release quality
Yes
post-release quality
programmer productivity
customer satisfaction
N/A
team morale

11. Conclusions XP “successful” for two small, co-located teams
“It was the best of times, it was the worst of times.”
OK, no “worst times” detected.
Though some may still yearn for the structure of plan-driven methods a personal, cultural thing.