I would weaken that statement a little bit. The numbers are fine. The *conclusions* are bullshit.
As an example, there were 13 systems total over a million lines of code. Yet the authors drew grand conclusions. The number 13 isn't bullshit. It just isn't large enough to draw really strong conclusions. On Mon, May 13, 2013 at 11:30 PM, Ulrich Stärk <u...@apache.org> wrote: > Much more interesting is the statistical analysis in the comments (which > is missing from the main > article) that concludes: these numbers are bullshit. > > Uli > > On 13.05.2013 21:24, janI wrote: > > http://www.wired.com/wiredenterprise/2013/05/coverity-report/ > > >
