Download phill2006Welcome2Promise_slides

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Promise 2006
Gary Boetticher, Ph.D.
Co-chair
[email protected]
Tim Menzies, Ph.D.
Co-chair
[email protected]
1
Past, future
• 2004:
– Predictive Software Modeling, Chicago,
• Jelber Sayyad Shirabad, Tim Lethbridge,
Stan Matwin
• 2005:
– Promise data repository on-line
• http://promise.site.uottawa.ca/
SERepository/
– Promise1, St Louis (with ICSE)
• Tim Menzies
& Jelber
– IEEE Software Special Issue (Nov’05)
• The Promise of Public Software Engineering
Data Repositories
• Guest Editor: Bojan Cukic
• 2006:
– Promise2, Philadelphia (with ICSM)
• Gary Boetticher & Tim & ICSM
• 2007:
– Promise3, Minnesota (with ?ICSE)
• You? & Gary & Tim
• ICSE workshop proposals due Oct 6
2
Some details
• Last thing today: discussion
– Are we living up to the
promise of PROMISE?
– Should there be a PROMISE 2007?
• 2006 Proceedings
– On CD, at web site
– Authors retain copyright
• Sorry, no special issue this year
– Submission base needs to be wider
– Promise’06 & Promise’07
authors can submit
• A.M. & P.M. coffee:
– With ICSM (with thanks)
• Dinner tonight
– We’ll buy. Where to go?
• Receipts: see Tim M.
3
What makes
PROMISE different?
• Put up or shut up
– If you conclude X, give others enough
information to check X
• SE experiments
– Repeatable, refutable, improvable
• Promise repository
– http://promise.site.uottawa.ca/
SERepository/
– 2004:
• “It’ll never work”- Lionel Briand,
– 2006
• Currently, 2 dozen data sets
4
Challenges
• Where is the science?
– Currently, no repetition
• Where are the new technologies?
–
–
–
–
–
Text mining
Feature subset selection
Bayes nets
SVDDs
Etc
• Where are the landmark results?
– Stop sweating the petty things
• E.g. 2% mean accuracy
improvements
– Report significant improvements
over older work
5
Are we getting
the big picture?
• We are software engineers. Practitioners
• Where are studies on the feedback loop?
• What are the impacts of our learned
theories on:
– An evolving model?
– The organization?
6
7
Invited Speaker
• Predictive Models in Software
Engineering: State-of-the-Art,
Needs, and Challenges
– Lionel Briand Ph.D.
• Carleton University Software Quality
Engineering Laboratory,
Ottawa, Canada
– Canada Research Chair in Software
Quality Engineering,
– co-editor in chief of Empirical Software
Engineering: An International Journal
– “I am now a proud Canadian (or
should I say Canadien) though I still
have a weakness for pungent
cheeses. (well, nobody is perfect and I
did not have to give it up to become a
"canuck".)”
8