Gabor Szabo schreef: > Two days ago or so I posted a blog entry about the new CPANTS metrics. > http://szabgab.com/blog/2008/06/1212827982.html
What is the purpose of the CPANTS Kwalitee metrics? The most important reason for these kinds of ratings is to help users find the best modules for a certain task. Kwalitee is almost useless for this, because it only measures some sort of compliance with techynical best practices, while leaving unanswered questions about usefulness, completeness, etc. Cpanratings is much better at informing users of the quality of CPAN modules. So in my opinion, the main value of the Kwalitee metrics is not for the users, but for the authors of a module. Those of us who play the "Kwalitee game" will change and improve our modules to increase our Kwalitee level. This works as long as we, the authors, can influence the Kwalitee score of our modules. If too many of the Kwalitee points depend on external factors, Kwalitee will cease to be an (somewhat) accurate measure of the amount of work an author is willing to put into a module, and it will no longer be a matter of pride for authors to get a high (or even perfect) score. Personally, if the debian related Kwalitee measures are accepted as core indicators, I will no longer be interested in raising the Kwalitee of my modules. This means I would lose part of my motivation to improve my modules. Depending on whether you think Kwalitee correlates with quality, this would be a pity. Eugene
