COQUALMO Working Group Presentation Presenter: John D. Powell 1

Download Report

Transcript COQUALMO Working Group Presentation Presenter: John D. Powell 1

COQUALMO Working Group
Presentation
Presenter: John D. Powell
1
Participants
• Moderator: Sunita
Chulani-IBM
• Scribes: John Powell-JPL
& Keun Lee-USC
• Floating Participant:
Barry Boehm-USC
• Participants
• Michael Crowley-Motorola
• Kimberly Dobson-Motorola
•
•
•
•
•
•
•
•
Jongmoon Baik-Motorola
Linda Brooks-TRW
Barbara Hirsh-Motorola
George Huling-LASPIN
Jung-Won Park-USC-CSE
John Serino-BAE Systems
Dick Stutzke-SAIC
Nancy EicklemanMotorola
2
Outline
•
•
•
•
Preliminary Presentation
Brainstorming Session
Identification of Research Issues
Prioritization of Research Topics
3
Preliminary Presentations
• Presentation for those who were unfamiliar
with
– COQUALMO
– Orthogonal Defect Classification (ODC)
4
The COQUALMO Model
Baseline Defect
Intro Rates/Ksloc
Rqts
Rqts
Rqts
10
20
30
DI_DriverR,1
COCOMO II
Cost Drivers
DI_DriverR,1
QAF = DMR
Analysis
Tools Rating
Peer Reviews
Rating
Test Thoroughness
and Tool Rating
10*
DAFR
20*
DAFD
30*
DAFC
DR_DriverR,1
DR_DriverR,2
DR_DriverR,3
5
ODC
• Defect Classification, Trigger activities
• Signatures of defect rates
• Comparison between projects and historical
signatures
• Take action based on these differences
6
Brainstorm
• ODC
– Tailorable to different environments
• Modeling for Spiral V. Waterfall Lifecycle
– Modeling for Spiral life cycle is harder
• Finding the Right Kind of Defect V
Counting Defects
– Implication for Risk Exposure Assessment
7
Brainstorm (cont’)
• Temporal information about defects may be
needed
– Account for timing of activities
– Implication for cost
• Granularity of COQUALMO Calibration
– Allow for USC to Collect Data in the form
affiliates collect it and make use of it.
8
Brainstorming (cont’)
• Can we compress 8-9 defect categories into
3-4 categories?
• Build the capacity into the model to allow
for changes in its classification scheme
9
Identifying Issues/Research
Topics
1.
2.
3.
4.
5.
6.
7.
Full ODC Categorization
Cost/Benefit Analysis
Tailorability to waterfall, incremental model
Tailorability to Non-ODC Categories
Eliminating obstacles to data contribution
Reliability/MTBF – Product Model
Process Model
10
Prioritizing Research Topics
11
1. Full ODC Categorization
Importance
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
Difficulty
12
2. Cost/Benefit Analysis
Importance
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
Difficulty
13
3. Tailorability to waterfall,
incremental model
Importance
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
Difficulty
14
4. Tailorability to Non-ODC
Importance
Categories
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
Difficulty
15
5. Eliminating obstacles to data
contribution
Importance
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
Difficulty
16
6. Reliability/MTBF – Product
Model
Importance
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
17
Difficulty
7. Process Model
Importance
10
9
8
7
6
5
4
3
2
1
0
0
1
2
3
4
5
6
7
8
9
10
18
Difficulty