| || |
How to derive defect density?
Hello Fellow Testers,
As part of testing deliverables, we are required give a defect density.
I searched Google and most of the search results I got talk about defect density in terms of (No. of Defects / KLOC).
I searched qaforums, but didn't get any results meeting my requirements.
But in Object oriented languages like Java, it would be difficult to arrive at (K)LOC, as there would be more lines of comments than actual code. OO languages, there would be more instances of methods...etc being called etc..etc.
Also the (No.of Defects/KLOC) formulae was defined long long time, when COBOL & Mainframe systems were in vogue, so applying the same now... I am apprehensive, if this would be correct.
I would like to request you, testing gurus to guide me on how to derive a defect density with regard to Java?
Looking forward to your guidance.
Thanks & Regards
Re: How to derive defect density?
(interview question answer) [img]/images/graemlins/smirk.gif[/img]
Defect Density is a coding/programming term and is measured by lines of code. You may see it expressed as xxx defects / yyyK LOC. Then your organization would define what these measurements mean in terms of quality:
100 Defect = Excellent Quality
1000 Defects = Good Quality
10000 Defects = Marginal Quality
100000 Defects = Poor Quality
(get this task done by tomorrow answer) [img]/images/graemlins/wink.gif[/img]
I would think the QA team would be expected to report defect density from a functional perspective. This will require some defect tracking maintenance if you do not have a method in place to categorize your defects by functionality. If your defect tracking tool cannot manage this for you, this can be accomplished in MS Excel. Once the functional categories are in place, you will better understand which functionality contains the largest number of defects and could be said to have the highest defect density.