Ninety Day CPAN De-List Proposal

[InfiniteSilence]

I find a module. I like it. It has a problem. It is a simple problem and so maybe I write a fix for it or perhaps there is an obvious thing that it should do. I find the e-mail address in the POD. I send and e-mail an it gets bounced back because that e-mail address is no good.

I think that CPAN itself should send a confirmatory e-mail to EVERY module author with a code and if that author does not type in that code in some screen the module should AUTOMATICALLY be pushed to some back archive, effectively de-listing it from CPAN.

Note: I am not saying that people have to change their modules due to requests. I am saying that I should *never* get a bounced back e-mail. To me that means the module should no longer be in CPAN.

Share your thoughts, please.

[pryrt]

Share your thoughts, please

So if I change email provider, but haven't yet gotten around to updating my email at CPAN (a service where I don't expect frequent, if any, emails to come from) you want CPAN to delist all my modules?

So if I have linked my CPAN distributions to an issues-tracker (either rt.cpan.org or github issues or other appropriate tracker), and don't want to list a valid email on CPAN for spam harvesting, you want CPAN to delist all my modules?

And, by the way, I don't think you have to make your email public in your CPAN profile (though maybe if you do that, it just shows your cpan.org address? I know that's what I have publicly visible, but it's been a while since I set it up); and I don't remember seeing an email-the-author link on most distribution pages. Are you talking about the emails that show up in the AUTHOR section of the POD? If that, so you expect that an author will make a new release of all their distributions if they change their email address, even when they don't have any code fixes?

And even if a module isn't being actively maintained, it can still be of great use to many people, so delisting it just because you cannot get an email through to the author seems rather an overreaction.

All-in-all, your "solution" causes a lot more problems than you think it solves.

The first line of getting help or making suggestions is to look for the Issues link at the distribution page on metacpan.org. If there isn't such a link, see if the Repository link has an attached issues page.

And if the author is unresponsive, and you or someone else wants to be able to take over responsibility for that module, there is a well-defined process, as described in the "take over" section of the PAUSE documentation and in perlfaq7, rather than just arbitrarily hiding/removing their previous work from CPAN.

So, IMO, you are proposing a bad "solution" to a problem whose reasonable solution is already implemented and publicly documented.

---


edit 1: add parenthetical after reading marto's reply
edit 2: rephrase parenthetical

[hv]

So you find a module, and like it. But due to an email problem you think it should be delisted so that neither you nor anyone else can use it. It is unclear to me that anyone at all benefits from that.

There have been suggestions in the past to delist modules that never worked from the first time they were uploaded, and have not been updated in over a decade. Those suggestions have also been rejected by the PAUSE admins.

However there are mechanisms to become the maintainer or co-maintainer of a module whose original author no longer has the time or inclination to maintain it themselves; I believe in rare cases that can happen even when the original author cannot be reached to give their blessing - but it would take substantially more effort than "I tried emailing them and it bounced".

Update: I ++'d your node despite strongly disagreeing with the content, since I feel that the provoked thread is beneficial to PerlMonks.

[marto]

"I think that CPAN itself should send a confirmatory e-mail to EVERY module author with a code and if that author does not type in that code in some screen the module should AUTOMATICALLY be pushed to some back archive, effectively de-listing it from CPAN."

A valid email is required to setup a PAUSE account, your proposed solution doesn't seem workable to me. The address may be valid at the point of your proposed challenge, this of course may not remain true in perpetuity. I also don't agree with automated desisting of modules for this reason.

[hippo]

I concur with all the other replies so far. Delisting an entire dist just because of a (perhaps temporary) problem with delivering an email to the author seems unfeasibly draconian.

I find the e-mail address in the POD. I send and e-mail an it gets bounced back because that e-mail address is no good.

Then your next course of action should be to submit your fix to the issue tracker. In all honesty that should have been your first course of action rather than badgering the author on their email account.

[kcott]

G'day InfiniteSilence,

This seems like a knee-jerk reaction to me.

A more productive question might have contained content along these lines:

"I'm trying to contact X, the author of Y, whose email address is given as Z. I'm not having any success. Can anyone provide alternative contact details?"

I have seen many posts such as that in SoPW over the years. They mostly receive informative replies.

[SankoR]

You may not get a response in under 90 days regardless of whether or not the email bounces. Most of CPAN is unpaid and unsponsored work strictly done in our spare time and life can be hectic.

Does the module you're talking about have a public issue tracker? How long ago was the most recent dist uploaded? How long ago was the author's last upload of other modules? Jumping to module takeovers after a month and a half just over bounced emails seems drastic.

[Bod]

I am saying that I should *never* get a bounced back e-mail. To me that means the module should no longer be in CPAN

I feel your frustration...believe me, many companies with paid staff have the same frustration and have yet to solve the issue fully.

However, your proposal forgets that everyone submitting modules to CPAN does it voluntarily. If I thought my modules might get de-listed because of a problem with my mail delivery, I might be less inclined to put them there in the first place.

Then there are the modules that are useful but don't need much updating. I dare say some authors will have passed away, yet their modules are still being used as a lasting legacy to their ingenuity in creating the module and generosity in sharing it. And what of module dependencies? If you publish module ABC that depends on module XYZ and XYZ gets delisted...then what?

Thinking the other way, I do put my real email address in the Author section of POD in case CPAN's mail gets shut down one day...

[harangzsolt33]

am saying that I should *never* get a bounced back e-mail. To me that means the module should no longer be in CPAN.

very bad idea.