Upload
chloe-gallagher
View
213
Download
0
Embed Size (px)
Citation preview
Slide 1
Experiment Framework
Results from the Working Groups
Revised Experiment Plan
People for Roles
Who signed up for what
Goals 2001/2 until next ISERN Meeting
Next steps
ISERN Distributed Experiment:
Action Session
Slide 2
Framework for the distributed experiment
• Industry Survey
– Descriptive analysis, no benchmark.
– Report characterizing types of responding organizations and their inspection processes.
• Benchmark an inspection technique
– Pick some inspection variant and study it across different contexts.
• Controlled experiment
– Get very accurate data about the improvement due to some inspection variant.
• Shadow Experiments: To lower risk before investing in Level 2 or Level 3 for industrial collaborators.
– ISERN will find a university course that would be willing to perform a first run of the same study.
Slide 3
Results Working Group Survey
• Discussion of Survey Goal
– Original Goal: Characterize context for inspection experiments (and collect some data also from organizations which do not currently use inspections)
– Input from Working Group: Also… ask survey respondents (who will be development team members rather than organizations) what they do for defect reduction early in development (as many development teams do not use formal inspections).
• Inspection survey questionnaires from Lund
– Were looked through and there were no objections to use them as basis for the development of survey instruments.
• Next steps
– Develop questionnaires for context variables of experiment
– Option to plan a larger survey on early defect reduction approaches (based on experiment context survey?).
Slide 4
Goals for the technique to be benchmarked
• Require evidence of its likely benefit
• Should be widely applicable (maximize potential pool of participants)
• Some version should be able to be taught in a “reasonable” time
• Should be of genuine interest to target audience
• Results should be actionable
Slide 5
Results Working Group Experiment Design
• Decision on experiment topic:– Design Inspection (UML?)
• Target business areas for experiment:– Embedded systems, banking, insurance, …
• Catch interest: package as training– Possibly use situated learning? – 2 days– Available on web/streaming video– Pricing policy?– Training changes the document to be inspected?
• Ethical issues: need common NDA
• Experiences so far with design inspection
• Material available from past experiments
Slide 6
Who signed up for what?
• Sign-up status:
– Survey (without experiment): 12 X, 3 ?
– Experiment (with context survey): 6 X, 11 ?
– Shadow experiment: 11 X, 2 ?
• If you have not signed up, and want to do so:
– Stefan Biffl has the list.
Slide 7
Organization: People for roles
• Librarian: Marcus Ciolkowski.
• Working group chairs
– Experimentation design: Forrest Shull
– Survey design: Ross Jefferey
– Meta-analysis: ?? (plans data analysis)
• Steering Committee: contributors to experiment
• Experiment Preparation Committee: authors of experiment documentation
• Experiment Coordinator: knows experiment status
• QA for Data Repository: checks incoming data
Slide 8
Potential Steering Committee Members
• OORT experiments
– Shull, Travassos, Conradi
• Inspection
– Miller, Wohlin, Maldonado, Succi
– Wood, Dunsmore, Karlsson
– Biffl, Ciolkowski
• Esernet
• If you want to take part in the Steering Committee, talk to Stefan Biffl.
Slide 9
Goals and Next Steps
• Next steps
– Make package of currently available material for design inspection (in English, for translation by local researchers as they see fit).
– Send package out to experiment group for review.
• Time table
– End of September 2001: 1st version of package and survey instrument.
– X-mas 2001: Final version of package and survey instrument.
• Goals for 2001/2 until next ISERN Meeting
– Conduct and evaluate survey with inspection focus.
– Conduct some experiments.
– Conduct meta-analysis on experiment results and data from past experiments.
Slide 10
Server for Discussion
• http://asercweb.ee.ualberta.ca:8000/ISERN
• Login: name
• Password: Lastname