Managing Testing and Integration Through Quantitative Data

How can you possibly estimate how much time you need to spend on testing and integration? We all hope for the best but realistically how long should it take us for a particular project? And when we are testing a module how do we know if it would be cheaper to continue to test and fix the module or if it would be cheaper just to re-design and re-write it?

Ellen George of Software Six Sigma has written an article in the Fall 2004 issue of Methods & Tools called “Managing Your Way through the Integration and Test Black Hole”. In her article she covers how the effort needed to test and integrate a system. How to see if your codes quality is meeting expectations or if a red flag should go up. She also shows you how to determine early on during the testing of a module if it should just be thrown out and re-written.

In her articles she mentions that 70% of defects found are in 20% of the modules and 50% of defects found are in 10% of the modules. It would of been interesting if the number of defects in a module were correlated with the complexity of the module. My assumption is that the more complex a module or the higher risk modules would have a higher defect rate and even if they were re-written the defect rate would still be similar.

written by: John Kopanas

Related Links:


Help Support Us By Visiting Our Sponsor

10% of our revenue from these ads go towards scholarships!


Comments


Post Your Comments





(note: not displayed)


(note: displayed with comment)

CUSEC 2005

CUSEC is the Canadian Undergraduate Software Engineering Conference created to promote software engineering in canada at the undergraduate level. CUSEC 2005 is being held this year in Ottawa, Canada.

http://2005.cusec.ca

Monthly Newsletter

Sign-up to our monthly newsletter so you can keep up with the latest articles on software engineering and our conference (CUSEC).

sign-up now