[pm-h] NeilB commentary on CPAN

[G. Wade Johnson]

On Mon, 11 Aug 2014 03:45:53 -0700
"Michael R. Davis" <mrdvt92 at yahoo.com> wrote:

> > http://neilb.org/2014/08/10/prepared-to-put-up-with.html
> > Makes you think about a central part of the Perl ecosystem.
> 
> G. Wade,
> I think if CPAN had a few features then it would be better.  

Hi Mike,

One of the great things about CPAN is that it is maintained by the
community. (That can also be one of it's downsides.)

The code for metacpan.org is on github as projects under
https://github.com/CPAN-API

This allows individuals or groups with an idea to contribute and
improve the interface.

> 1) I use this distribution please keep it updated.

I have used a cron job in the past to try to do this. I eventually gave
up on the idea thanks to one too many breaking change in a
module.<shrug/>

> 2) Don't use this distribution use this newer, bigger, better package
> instead. 

That's a really good idea to be able to add to metadata about a module.
This would allow an author to easily let people know if they want the
module to be replaced.

Unlike the previous search.cpan.org, metacpan.org supplies a RESTful
API that you can use to query CPAN for information you need for your
environment.

> So, the packages with the most "use" could be filtered from the rest
> more easily. 

I've seen a fair amount of talk about this over the years. 

> Frankly, I've uploaded some packages and have no idea whatsoever if
> anyone is using them.  So, do they need to stay up on CPAN even
> though I no longer use them? Thanks,

Modules uploaded to CPAN stay there until the author removes them. I've
been surprised a time or two when I see that someone is using a module
that I wrote and had basically forgotten.

I guess the short version is that you could either make changes to
MetaCPAN yourself, get a group together to look at it, or at least post
a suggestion on the relevant MetaCPAN project. It's definitely possible
to extend MetaCPAN to do more.

G. Wade
-- 
Virtual is when it's not but it looks like it is and transparent is
when it is but it looks like it isn't.                     -- Rick
Hoselton