All the Perl that's Practical to Extract and Report

The CPANTS Game

davorg on 2006-06-26T08:18:21

Repeat after me. "CPANTS is _not_ a game. CPANTS is _not_ a game".

So I spent a few evenings tidying up all of my CPAN distributions. They now all have the highest kwalitee that I can give them. The only points missing are the ones for your module being used as a pre-requisite by someone else.

As a result, I leapt quite a long way up the CPANTS leaderboard. Which made me very happy.

But now, of course, every new module I release can only have a maximum kwalitee of 17 (as a new module won't be a prereq for another module). So releasing new modules can only bring my average kwalitee down. Any increases in my score are in the hands of other people.

All of which could possibly act as a disincentive for releasing new modules. And that's why I now have to keep reminding myself that "CPANTS is _not_ a game..."

Re: CPANTS

domm on 2006-06-26T08:47:29

I'm looking for a solution to this problem, but haven't found one yet. Simply adding the number of dists to the rating would distort the stats in favour of people with many dists. So I/we need to come up aith a way so that adding a new dist doesn't lower ones rating. Ideas welcome!

Re: CPANTS

jplindstrom on 2006-06-26T11:56:51

Well, what's the kwalitee score used for?

If it's used to determine the kwalitee of a module, it's useful to know whether someone else uses the module (actually, more interesting than some other stats).

But the kwalitee of an author? That's just weird and not very useful except as a second-order measure of whether a module is good. So that number isn't very useful anyway, except as a vanity thing.

Either you remove it altogether since people tend to misuse it.

Or you game the number so that it steers behaviour in the right direction, i.e. you remove the point for "being a dependency" when you summarize the kwalitee score for an author.

Re: CPANTS

jhi on 2006-06-27T05:57:21

Engage the prereq rule only after the module has been released for a year? (Yes, this would mean that the CPANTS rating should probably by a percentage instead of an integer.)

L'enfers, c'est les autres

ferreira on 2006-06-26T11:42:03

Just how Sartre's character said. I think the is_prereq rule tests whether the exercise of laziness/patience/hubris of a Perl programmer was successful to reach the world. What is good for if nobody saw all your care to not be bothered for features people would want (anticipating them with your laziness), all the patience to write docs and flawless code, and the fine tuning to assure your hubris. I think there's nothing wrong with this CPANTS rule which at least gives a (possibly distorted) acknowledge of the community for your work. Well, of course there's the kind of modules which have not been written to be prerequisite of others, like standalone scripts, complete frameworks, etc.