There is/was something in the works. CPANTS (Comprehensive Perl Archive Network Testing Service) has been talked about for some time and work appears to have begun on it.

• CPANTS Synopsis & Discussion
• Schwern's YAPC 2001 slides discussing CPANTS
• CPANTS' Source Forge Page (Nothing of value submitted yet)

If you want to help, there's a developer list you can subscribe to here.

As part of this there is meant to be a karma rating for authors and modules.

At YAPC::Eu::2001 I signed up for a working group to implement a karma. OeufMayo and BooK were also involved. I think that we all felt that the PerlMonks system could be used in some way. However, I haven't written any code and judging by the mailing list no-one else has either.

Nevertheless, it is a good idea and it should be followed through.

Does anyone have any opinions about the suitablity of the Slash, PerlMonks or Everything engines for this task?

--
John.

The other downside is that as soon as you see a message like:

If you like this module, visit vote.cpan.org and vote for it.
[download]

And then CPAN will turn into a great big vote-whoring exercise and before you know it, someone will write Natalie::Portman and Hot::Grits.

counting downloads wouldn't work, though, because of all those pesky mirrors

This is true if FTP downloads are counted. However CPAN.pm and PPM could be modified to notify a central server every time a user installs or updates a module. Of course, not everyone uses CPAN.pm or PPM.

Also, this may be considered undesirable from a privacy standpoint.

The only problem with using this as a metric is that it doesn't measure whether or not you ever used it after you installed it, or what your opinion of it was. I've downloaded *tons* or modules from CPAN, but that doesn't mean that I would rate them all the same. Sometimes I'll download one thinking it may do what I need, but upon closer examination I find that it doesn't (or worse, doesn't even work). I wouldn't want my constant useage of HTML::Template to be rated the same as my curiosity download of (the previously mentioned) Natalie::Portman.

I don't think counting downloads alone in any fashion is going to provide the kind of info to be really helpful. However, it would be good to have it along side a 'vote' based system, so that you could tell what percentage of the people that downloaded it rated it well.

/\/\averick
perl -l -e "eval pack('h*','072796e6470272f2c5f2c5166756279636b672');"

If that's any help once upon a time I designed a Module Review Report. It might be a starting point to design an application that would let people give their opinion on modules.

A general index listing categorized modules (that's easy, the namespace provides a convenient category) and the summary could link to the detailed report (which could be sorted on the various fields).

BTW I think it is important that comment authors be listed, as it prevents "ballot stuffing" and gives readers a way to evaluate the credibility of the rating.

• Not based on downloads (although that could be statistic for the curious)
• Non-anonymous voting
• Reviews/comments/notes
• Natalie::Portman gets +10,000 bonus points

At the moment the review section is... just an idea, there is no code whatsoever to support it. I would think that in order to work it would have to be part of PerlMonks, just so users can be authenticated by the site.

I (or anybody with half a clue and more time on their hands!) can write a quick proto, just to demonstrate the concept and get the UI right, but then the maintainers would have to integrate it in the site.

Well, I've got the next two days off work so I might have a quick go. Maybe this is the moment to learn about writing nodeballs B-)

Authentication-wise, we can can always have our own list of users, although admittedly a Perl Monks tie-in would be great.

Where's the "Summon a Perl Monks Maintainer" button when you need it?