Talk:Software inspection
From Wikipedia, the free encyclopedia
Contents |
This article needs to be combined with code review. --Stevietheman
- Nevermind. There appears to be some differentiation. --Stevietheman 18:00, 28 Dec 2004 (UTC)
[edit] software inspection as peer review
As identified in the article, software inspection is recognised as a form of *peer review* in the software world, but there are at least two deficiencies that should be remedied.
1. The Wikipedia entry on "peer review" considers only the academic publication use of the term, not the software enginering use.
2. The brief discussion of "peer review" in the "software inspection" article omits "technical review" as a type of formal peer review (see IEEE 1028), and omits informal peer review (group review, "buddy check") altogether.
See further comments under the "peer review" article.
--Donmillion 22:07, 28 April 2006 (UTC)
I have edited / created articles to form the following hierarchy of cross-linked entries, drawing on IEEE Std. 1028-1997 (Software Reviews) for conceptual normalization:
|--|- software peer review
|--|--|- software walkthrough
|--|--|- software technical review
|--|--|- software inspection
|--|- software management review
|--|- software audit review
However, the article on software inspection needs to be completely recast to conform to the general model I have established for walkthrough, technical review, management review, and audit review. It should also be strengthened to identify the unique features of inspection (emphasising training, cost-effectiveness, process and skill improvement, the use of a trained Moderator, the use of oracles, rigorous numeric entry and exit criteria, and collection and analysis of metrics). I am reluctant at present to completely rewrite so extensive an article as the present one, but feel constrained to remark that it illustrates the reason why Fagan has repudiated the term, "software inspection" (because his very clear and precisely-defined process had been "hijacked" by people and organisations who were not prepared to apply it with the necessary rigour, and compromised on many of the unique features outlined above).
--Donmillion 22:07, 28 April 2006 (UTC)
The article says that "The goal of the inspection is to identify and repair defects." As originally defined and commonly practiced I don't think repair is a primary goal. The problem is it's too easy for the meeting to get bogged down with a debate over how to fix a defect. If someone mentions a fix and everyone quickly agrees then no harm no foul, but any debate will take time away from the task at hand which is finding defects. --Eddie
