On Mon, 2008-10-27 at 19:17 +0100, Salve J Nilsen wrote: > chromatic wrote: > > On Monday 27 October 2008 10:45:46 Salve J Nilsen wrote: > > > >>> Remember, this is not a project designed only to say "This code sucks." > >>> Its intent is to encourage people to improve their code. My code > >>> doesn't magically get better when someone finds a bug. It magically > >>> gets better when someone *fixes* a bug. > >> > >> One is a prerequisite of the other. You have to have some indication that > >> a bug exists before you can fix it (let's ignore "accidental bugfixes" for > >> now.) So unless you live in a bubble all by yourself, this list will at > >> the very least increase the likelyhood of you learning about (in this case > >> Kwalitee) bugs. > > > > A public hall of shame that several people on the Perl-QA mailing list did > > not know about has a very marginal effect on increasing the likelihood of > > learning about a problem. I'm not a statistician, so I can confidently say > > that the chance of that occurring is non-zero. (Randomly stumbling across > > several billion web pages will *eventually* get you there.) > > We're still talking about a marketing/visibility bug here. Don't you agree > it's > better to fix that instead?
You think that it is important that the CPANTS team makes sure that everyone knows there is a web page dedicated to shaming developers? > > > >> This is a good thing. Especially if the scale we're measuring the code is > >> sensible, well thought out and relevant. If your ego gets a bruising, too > >> bad. The code Kwalitee is more important. > > > > Heaping random, unsolicited, and public abuse on contributors is a fantastic > > way to make sure there are no Kwalitee programs -- in the sense that > > abusing contributors is a great way to make sure that there are no more > > contributors. > > There's nothing random or abusing here, just feedback on Kwalitee comparisons > between modules. If this feedback hurts your (or anyone elses) tender little > feelings, then too bad. A psychologist would remind you not to equate > critique > of your writings with critique of yourself. You are entitled to your opinion of course, but one doesn't need a psychologist to identify ad hominem. Antagonism doesn't breed quality software. If it is really a goal to increase quality, then a 'hall of shame' is counter-productive, the feelings of people on either side of the issue notwithstanding. If the hall of shame really is 'Kwalitee comparisons between modules', it doesn't require a page title of 'hall of shame' and it would by your definition need to include all modules under test rather than a subset deemed worthy of shame. Some person or committee got together and actually decided to spend time and energy putting this together. I would have loved to have been a fly on the wall for that conversation. "Maybe if we produce a hall of shame, the n authors on this list will finally include a README.txt and a META.yml in their distribution! Then we'll achieve true 'Kwalitee'!" (I think the correct pronunciation for this should be qvalitee....) Has anyone actually produced metrics regarding the movement caused by CPANTS? ie; When the first 'hall of shame' was produced, how many authors distributions were on that list and how many moved off the list? What metrics exist to show the effectiveness of the various CPANTS initiatives? For example, the 'extractable' criteria shows zero non-compliant distributions. Was this ever >0? If not, wouldn't it be reasonable to assume that a test for 'extractable' is a complete waste of time and energy? > > > - Salve >
