Tree

Navigation ...

lh_home.png
lh_personal.png

Scientific Writing ...

lh_AA.png
lh_AC.png
lh_BE.png
lh_CS.png
lh_DA.png
lh_ED.png
lh_GE.png
lh_GP.png
lh_LW.png
lh_MA.png
lh_ME.png
lh_PR.png
lh_RE.png
lh_SA.png
lh_SF.png

Other Writing ...

lh_AR.png
lh_ES.png
lh_MU.png
lh_SP.png

Reference details

Author(s) Year Title Reference View/Download

Les Hatton

2005o

Predicting the total number of faults using parallel code inspections

WEBInspect2005.pdf

Synopsis and invited feedback

Peer review is important and acquiring competent reviewers is becoming a major problem for the journals today so I will be very happy to include constructive comment (positive or negative) with acknowledgement. If I am not competent to judge your commentary I will try and find somebody who is.

If you would like to provide feedback just e-mail me here.

Synopsis Invited Feedback Importance (/10, author rated :-) )
Shows two interesting things. a) Parallel code inspections can be used to predict the total number of remaining faults in a software system (during which I rediscovered a technique which I now learn is well known in fisheries, Capture-Recapture), and b) Using 22 parallel 2 person inspection teams, it is shown that independence CANNOT be rejected. This latter result is really surprising as it would be expected that checklists would force some kind of obvious dependence. I went on from this to study it in detail on a much larger scale as shown in the links field. The effect is still there:- checklists appear to have no significant effect on defect detection quality.None yet9

Related links

Related papers and links

http://www.leshatton.org/checklists_in_code_inspections.html


Auto-generated: $Revision: 1.58 $, $Date: 2016/01/03 19:07:28 $, Copyright Les Hatton 2001-