Beefy Boxes and Bandwidth Generously Provided by pair Networks
go ahead... be a heretic

Re^5: JETTERO tries to take over Net::IMAP::Simple on PAUSE (FUD)

by tye (Sage)
on May 19, 2009 at 04:05 UTC ( #764819=note: print w/replies, xml ) Need Help??

in reply to Re^4: JETTERO tries to take over Net::IMAP::Simple on PAUSE
in thread JETTERO tries to take over Net::IMAP::Simple on PAUSE

CPAN is already a mess. What we have now is 10 bugs not fixed and not reported because 10 people know darn well that the 80% of the time the only way to get a bug fixed is to take on the role of "stalker" in hopes of getting the lovely prize of "now you get to be the one that people are supposed to stalk", 4 different versions of each module with slightly different names and significantly different implementations and feature sets, none of which are getting improvements applied particularly well (most of the time) and none of which is the clear best choice just in terms of features. Short spurts of one person having the time to work on a module and then their life taking a turn and the module sitting untouched for years until somebody is actually perverse enough to jump through the hoop of "play stalker to the last person who touched it" for 6 months in order to be allowed to contribute to it (meanwhile, the module has been "forked" 3 times by people just starting over with a slightly different name).

Freedom and encouraging collaboration is scary until you do it. I'm a bit shocked that so many in the Perl community haven't overcome that fear. So CPAN will continue with its mini-empire building while a bunch of us put our modules on github and encourage collaboration and make it free for people to do the work of getting a module into shape for a new release and then releasing it. Eventually, freedom and collaboration will produce such better results that most of the CPAN mini-empires will just look sad.

There is no point in trying to explain this to the powers that run CPAN. They will just take it as a personal attack anyway. Instead, just act like the internet and "route around roadblocks" ("censorship" in the original quote).

Those who are invested in the closed-ownership and "we must have roadblocks to prevent chaos" will be scared and predict failure and may even be so close-minded as to feel threatened. Meanwhile, open collaboration will increase the pace of innovation such that it will stumble here and there but in the end will produce much better results.

I want to have the modules that I created and put onto CPAN in a "anybody can contribute" mode, but CPAN won't even allow me to do that. They don't support such "anarchy". Because we know that the concept of wikis were a passing fad that was simply doomed to a chaotic failure.

So don't waste your time trying to explain the better way. Just do it the better way and those who can deal will see the better way and use it. Those who fear contributions from others will cling to keeping their "ownership" (until they inevitably lose interest and their contribution withers and fades from memory).

Most open-source projects spend much of their lifespan being driven by a single developer. How well an open source project can support an orderly transition to the next community of contributors is a major factor in how long the project remains viable and useful. Allowing more Perl modules to much more seamlessly accept new contributors and transition to new owners will allow such modules to become mature (something that the vast majority of even useful modules on CPAN can't achieve). It will be a great benefit to Perl developers.

The policies of the CPAN services mean that CPAN is a huge collection of mostly just individual contributions where collaboration is the exception. We can do so, so much better. And we will!

- tye        

  • Comment on Re^5: JETTERO tries to take over Net::IMAP::Simple on PAUSE (FUD)

Replies are listed 'Best First'.
Re^6: JETTERO tries to take over Net::IMAP::Simple on PAUSE (FUD)
by jettero (Monsignor) on May 26, 2009 at 17:41 UTC
    You can assign as many co-maintainers as you like on pause. If you want it open, simply state that and then add people as they mail you. I really don't see what's so hard about that. In fact, it seems like about 12 lines of WWW::Mechanize code and you'd have an auto perminator on your web page. Then you could go to Hawaii permanently, end of story.

    I rather with cfaber would have done that. He seems to have lost the permissions to the two yokels who have it now... and they won't respond.

    If it were me, I would totally just assign co-maintainership, particularly if I didn't have time to work on it. And cfaber would have done so too.


      You can assign as many admins as you like on pause

      So I've heard it claimed. When I and another pointed out that the interface doesn't actually appear to support this, silence was the only response. (Re^6: Losing faith in CPAN - unresponsive module authors (ownership--))

      I really don't see what's so hard about that

      You presume I am even that active on CPAN. For most modules, that assumption is broken. I want to set up my modules so that others can contribute to them even after I get hit by a bus or become a Quarker, even for modules that nobody has yet expressed an interest in contributing to. And I want people to be able to contribute without there having to first be one person who has the unnatural motivation to play "stalker" for 6 months and then beg admins to deign to allow contribution. It would be even cooler if others could do the same. I really don't see what is so hard to understand about that.

      - tye        

        Yeah, the language in my post doesn't match your reply because I quickly found that you cannot give admin to others, only upload. I still say you could sort it out with WWW::Mechanize. I really don't think they'd object. I bet you could even get the github guys to add a service updater gadget.

        You could also fork pause on github and just write the patch. I wonder if andk would pull it in. There's two obstacles. Since people already use co-maint and expect it to mean a certain thing, you'd have to create a new perm name for co-primary or something. Second, you couldn't just give primary maint to more than one person because the module list only has a spot for one author name.

        ... hrm, I also definitely understand you. I don't know why you think I don't. I'm just thinking of ways around the problem. Believe me, I'm underwhelmed by this process so far -- on the other hand, I'd really hate it if I woke up one morning and they pulled the maintainership so I couldn't upload a new version... or worse, that someone else uploaded a version that changed the API all around and I had no way to block it -- not that I write anything anybody cares about. I'm just saying that some people definitely don't need this wiki stuff.

        I can't imagine trying to use LWP if anybody could just go in and change the meaning of get(). And while the vast majority of the edits would likely be well meaning, some wouldn't be. And some of those well meaning ones would be pretty bad. I'm sure you've seen the SoPW section here?


Re^6: JETTERO tries to take over Net::IMAP::Simple on PAUSE (FUD)
by rir (Vicar) on May 19, 2009 at 14:19 UTC
    I don't think there is so much fear around as you think.

    I want to have the modules that I created and put onto CPAN in a "anybody can contribute" mode, but CPAN won't even allow me to do that.

    This is confused. CPAN is the Comprehensive Perl Archive Network and is well named (Ha, how I hope I am not walking into another backronym. ). If you have the right, you can develop a codebase where, how, and with whom you like, then you can release it on CPAN.

    Licenses, established practice, and courtesy are arguments against summarily grabbing some author's code off of CPAN, reworking it, and putting it back on CPAN under the same name. It is not surprising that the CPAN admins are slow to declare code abandoned and then seize it.

    Be well,

      It is not surprising that the CPAN admins are slow to declare code abandoned and then seize it.

      It is not surprising that many choose not to contribute once they realize that contributing requires that "admins (slowly) declare code abandoned and then seize it". You nicely summarize the problem. Thank you.

      This is confused. [....] If you have the right, you can develop a codebase where, how, and with whom you like, then you can release it on CPAN.

      That seems to me to completely miss the point. Creation as the act of a single person is so very common. Maintenance is where collaboration is most useful. The ability to collaborate before submitting to the rules of the CPAN services is certainly not the problem, for multiple reasons. The need to thwart the roadblocks of CPAN services just to be able to collaborate should be seen as a problem, but, yes, we shall work around that problem.

      - tye        

        contributing requires the "admins ... declare code abandoned and the seize it"

        I assume I misunderstand you, because this strikes me as disingenuous at best. I would guess that any PAUSE author who requested that their work be taken off CPAN would have their request honored. It may provoke some comment and consternation, but given that an author may have legal issues regarding authorship I see no advantage in CPAN's noncompliance.

        I don't believe CPAN contributions require that "admins ... seize code" nor do I believe that the fear of such prevents contributions.

        CPAN is a distribution mechanism not a development tool. There is no reason to expect CPAN admins to jump to take on new tasks that have little to do with that for which they volunteered.

        When you release on CPAN, you don't have to stop developing the codebase where, how and with whom you like. I should have been clearer about the timing.

        There is nothing I see in the rules of the CPAN that you are complaining about; how is CPAN stopping you from setting up a collaborative development process for the modules in your care? You just seem to want CPAN to be something completely different than what it is.

        I like what you are proposing, I just don't see it having much to do with CPAN.

        Be well,

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://764819]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others making s'mores by the fire in the courtyard of the Monastery: (7)
As of 2021-05-16 18:23 GMT
Find Nodes?
    Voting Booth?
    Perl 7 will be out ...

    Results (152 votes). Check out past polls.