Stories
Slash Boxes
Comments
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

use Perl Log In

Log In

[ Create a new account ]

petdance (2468)

petdance
  andy@petdance.com
http://www.perlbuzz.com/
AOL IM: petdance (Add Buddy, Send Message)
Yahoo! ID: petdance (Add User, Send Message)
Jabber: petdance@gmail.com

I'm Andy Lester, and I like to test stuff. I also write for the Perl Journal, and do tech edits on books. Sometimes I write code, too.

Journal of petdance (2468)

Wednesday August 18, 2004
04:58 PM

Let's eliminate the Module List

[ #20456 ]
I propose eliminating the Long Module List. I'm talking about http://www.cpan.org/modules/00modlist.long.html (2998 modules), not http://www.cpan.org/modules/01modules.index.html (6800 modules).
  • It's no longer relevant.
    Way back when, it was cool to have a single readable source of information. With search.cpan.org, it's just not necessary any more. The list gives two aims:

    * FOR DEVELOPERS: To change duplication of effort into cooperation.
    * FOR USERS: To quickly locate existing software which can be reused.

    Both are addressed, and very effectively, by search.cpan.org.

  • Few people look at it.
    Looking at pair.com's mirror logs, I see that since Jan 2003, downloads of 00mod* have averaged fewer than five per month. Per month, not per day. Pair is not a lightly-used mirror, either. They served up 615K distros for July 2004. Five out of 615,000 is close enough to zero for me.
  • Inclusion on the list is effectively arbitrary.
    It doesn't mean anything to have a module on that list. It's certainly not a stamp of quality. I don't mean to ignite the debate over whether there should be some "Perl Approved CPAN module" apparatus should exist; only that inclusion on the Module List is not it.
  • The resources used could be better used elsewhere.
    There's significant amount of human time and machine resources that go into maintaining the Long Module List. For that matter, it's a waste of developer time proposing inclusion on a list that nobody looks at.
  • search.cpan.org browsing is misleading
    Browsing search.cpan.org gives the user the impression that he or she is browsing all modules on the CPAN. This is not the case. The 26 categories don't make sense any more, anyway.

The one bit of value that I see in this process is where Graham looks at submissions that people have sent in and, if something seems like it's duplicate effort, tries to redirect the author to reduce the duplication. (http://www.nntp.perl.org/group/perl.modules/34207) Unfortunately, that requires the author to submit a proposal for inclusion, and since fewer than half of the authors submit the modules, it's hardly a complete filter.

I welcome your thoughts. How can we capture the good part of the module list (the human filtering), and remove the obsoleted infrastructure?

xoxo,
Andy

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More | Login | Reply
Loading... please wait.
  • doesn't mean it isn't used or unimportant. The main menu on search is automatically generated from that list and it used to be as close to a list of publicly deemed worthy modules as one could get. People don't need to download the list that frequently anymore, they just use the index on search.

    And pair.com is the mirror that serves www.cpan.org.

    • I understand that the main menu is automatically generated from the list. See bullet point "search.cpan.org browsing is misleading".

      I also understand that "it used to be as close to a list of publicly deemed worthy modules as one could get", but that's no longer the case. Any ol' module gets on there, and plenty of good ones are ignored. I think it's safe to say that any one of Simon Cozens' modules would be "publicly deemed worthy", but there are only two of his 100+ on there. Ingy, 5. Schwern, 11.

      --

      --
      xoa

      • Not any module gets on the list. I don't think we approve even half of the submissions.

        I'm not sure any of Simon's modules qualify as anything. Wasn't his hundredth module only POD just so he could say that he had a hundred modules?

        I use the Module List quite a bit to help people name their modules or work with other people with similar modules. It does not some maintenance though.
        • Sure, and there are a couple of Bundle:: modules in there as well. That doesn't diminish the quality of the other 90+ that are worth putting out there. The point is that it's far from complete, or accurate. If we want to have a way to say "These are good, kwalitee modules", then let's do that. The Module List is no longer the way to do that.
          --

          --
          xoa

      • Not accurate because, in your opinion, there are too few of your friends modules? There are too many modules on CPAN and too little space in something like that list. It is intentionally narrow and small.

        It's accurate, but perhaps it's not as maintained as it once was since the modules list is impossible to manage and we're still waiting for an rt queue to help manage those.

        Aren't there bigger burning issues like boxers or briefs out there somewhere?

        • Not my friends, just people who have written good, solid modules. It's not like Simon and I are buddies, but he's still written good stuff. It's not that those modules are excluded, but rather that people aren't bothering to submit to it.

          Nobody's trying to take anything away from you, or any of the rest of the module list maintainers. I won't deny that there was great value in the past. That time was years ago, however. Let's start fresh with something new.

          --

          --
          xoa

          • Yes, that's what people said with Perl6, fresh and new...we can see how well that has worked out so far.

            CPAN is the one remaining thing that keeps a lot of people around perl and changes to it will be neither dramatic nor quick. Make a new list or propose something with an understanding of how things work, but there are better crusades to be had around these parts.

          • people aren't bothering to submit to it

            I think many just don't submit modules anymore because they know the list is not being updated.

    • Nitpicking... 6 hits in a month mean it's unused. Trying to call it "un-popular" or hinting that it is used, even though only 6 people a month use it, doesn't change this.
      As for being "important" - if it's misleading (e.g. people think it represents the entire module list) or drains resources away from more important endevours, than yes, I would call it "unimportant".

      It's past time somebody spoke out about this issue. I totaly agree- this list should be taken out to pasture.
  • I agree, 100%. I can't remember the last time I used the module list, but I find myself having to warn people away from it on a regular basis. I pity the novice that looks for a module there and concludes that CPAN can't help him when he inevitably comes up empty!

    I'm sure it could be repaired, but I don't think it's worth the effort. As it is it's worse than useless and should be discarded as soon as possible.

    -sam

    PS: And don't get me started about DSLIP codes!

    • As it is it's worse than useless

      I agree with this part...

      should be discarded as soon as possible

      ...but I totally disagree with this one! I think if we had an updated module list it would be to the benefit of our community!

      • No, this is why it's worse than useless.

        Not only does it do no good, but it does harm. One of the ways it does harm is by making us think that we have something reasonable in place, so we can "fix this later." Let's get rid of this system, which is not reasonable.

        If it then becomes apparent that we need a new system, we will feel more compelled to produce it. I imagine, though, that this need will /not/ become apparent immediately, and we will find that the registered modules list will be replaced grad
        --
        rjbs
        • I'm not saying "let's wait and fix this later"... I'm saying "let's fix this now!"

          I strongly believe that a long module list of that kind is something good for the community, as long as it is updated.

          Many of you may say you don't need it, as you use search.cpan.org instead... I don't use it either (in part because the list is outdated), but I know such a list would be a good thing to have.

          You can use search.cpan.org to search for specific modules, but others can use the long list to look at all those

      • I think if we had an updated module list it would be to the benefit of our community!

        That doesn't preclude us from throwing away the old and broken.

        --

        --
        xoa

        • No, it doesn't, but see my answer [perl.org] to rjbs above.

          If we have something old and broken that could be useful, why throw it away when we can rebuild it now?

  • I never quite understood what problem the modlist was meant to solve, much less whether it was solving it or not.
  • I too am in favour of getting rid of it. I have never used it, and I've never used the search.cpan.org directory either.

    In fact, as a beginner, years ago, the www.cpan.org interface confused and frustrated me. It was so hard to find anything! Manually browsing the site has been painfully difficult and inefficient for a long time. The module list does little to improve that.

    Nowadays, my only contact with CPAN is via search.cpan.org's query form and via CPAN.pm.

  • I welcome your thoughts. How can we capture the good part of the module list (the human filtering), and remove the obsoleted infrastructure?

    CPAN is pretty special, but it isn't unique. How do other sites solve this?
    Looking at CTAN [ctan.org] (on which CPAN was originally based), I see they allow browsing of the entire site contents, as well as search interfaces through several interfaces, including Google. So here's suggestion number 1:
    * Add a Google search!

    What else? I'm not sure. Maybe something like DMOZ [dmoz.org], bu