ACM Comm 2010 May Technical Perspective Automated Patching Techniques: The Fix is In (Notes)

From University
Jump to: navigation, search
"Beyond Total Capture: A Constructive Critique of Lifelogging" CACM May 2010

Technical Perspective Automated Patching Techniques: The Fix is In
by Mark Harman, p.108

Technical Perspective Automated Patching Techniques: The Fix is In

How is an OS that is actually a biological population, tested? Annual checkups? A walk-in clinic? How does a soft-life program know it is ill?

People

  1. Westley Weimer
  2. Stephanie Forrest
  3. Claire Le Goues
  4. ThanhVu Nguyen

Ideas

  1. Undecidable Questions Such as statement reachability. How is statement/method \(A\) reached for testing? Under what conditions is it invoked?
  2. Automated Bug Fix. Then regression testing all over again.
  3. Search Based Software Engineering (SBSE)[1]
  4. Every bug can't be fixed.
    1. What criteria is used in deciding which bug to fix and how much effort to spend fixing it.
  5. Zune Bug[2] was patched in a little over three minutes on standard equipment using SBSE. (Leap Year Bug).

References

  1. Automatic Program Repair with Evolutionary Computation by Westley Weimer, Stephanie Forrest, Claire Le Goues, and ThanhVu Nguyen. CACM May 2010 p. 109-116
  2. Arcuri, A. and Yao, X. A novel co-evolutionary approach to automatic software bug fixing. In Proceedings of the IEEE Congress on Evolutionary Computation (Hong Kong, June 1–6, 2008), 162–168.

Internal Links

Parent Article: Reading Notes