[dundee] Open-Source Code Quality Improving

[Robert Ladyman]

mickey.mouse at disney.com (Position: Cartoon Character) registered for me and 
allowed me to look at his copy. He told me that the use "Static Analyis" of 
defect density (so that they can compare projects) in lines of code and 
compare a scan of a project at one time with later scans. Their false positive 
(for defects) is about 14%. Mickey also said that the report gave roughly one 
defect per 3333 lines (0.3 near enough) but now it's down to one per 4000 
lines (0.25).

Mickey said that they don't look at function points, but at lines of code that 
aren't whitespace or comments.

Mickey also said that he couldn't comment on the statistical accuracy nor 
implications of this particular analysis as he is both a cartoon character and 
(notionally) also a fictional character. He did point out that allowing 
fictional characters to register does rather undermine the data gathering 
excercise but that he doesn't get much spam, however many people sign up as 
him.

Thanks Mickey.

> On 24 Sep 2009, at 11:39, gordon dunlop wrote:
> > 2009/9/24 Lee Hughes <toxicnaan at yahoo.co.uk>
> >
> >
> > obviously i believe you!!
> >
> >
> >   Lee, there is a link in the article to a page that lets people
> > download the report, so you can examine their findings.
> >
> > Gordon
> 
> It's this page, not all the links are correct
> 
> http://www.coverity.com/scan/
> 
> If anyone bothers to fill out the form, please let me know how they
> defined "quality".
> 
> Iain
> 

-- 

Robert Ladyman
File-Away Limited, 32 Church Street, Newtyle
Perthshire, PH12 8TZ SCOTLAND
Registered in Scotland, Company Number SC222086
Tel: +44 (0) 1828 898 158
Mobile: +44 (0) 7732 771 649
http://www.file-away.co.uk