Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > perl.module-authors > #3293 > unrolled thread

Where to report bugs in website 'search.cpan.org' ?

Back to article view | Back to perl.module-authors

Contents

  Where to report bugs in website 'search.cpan.org' ? law@tlinx.org ("Linda A. Walsh") - 2016-09-07 20:56 -0700
    Re: Where to report bugs in website 'search.cpan.org' ? olaf@wundersolutions.com (Olaf Alders) - 2016-09-09 08:57 -0400
      Re: Where to report bugs in website 'search.cpan.org' ? perl-diddler@tlinx.org ("L. A. Walsh") - 2016-09-10 04:01 -0700
        Re: Where to report bugs in website 'search.cpan.org' ? olaf@wundersolutions.com (Olaf Alders) - 2016-09-10 11:03 -0400
          Re: Where to report bugs in website 'search.cpan.org' ? perl-diddler@tlinx.org ("L. A. Walsh") - 2016-09-10 16:11 -0700
            Re: Where to report bugs in website 'search.cpan.org' ? olaf@wundersolutions.com (Olaf Alders) - 2016-09-10 20:10 -0400
              Re: Where to report bugs in website 'search.cpan.org' ? perl-diddler@tlinx.org ("L. A. Walsh") - 2016-09-10 18:18 -0700
    Re: Where to report bugs in website 'search.cpan.org' ? david@cantrell.org.uk (David Cantrell) - 2016-09-12 17:04 +0100

#3293 — Where to report bugs in website 'search.cpan.org' ?

Moving this question to module-authors as per David G.'s feedback.

I was looking for available mail-filtering, aka Milter, mods, and
came across Mail-Milter.

I noticed it has 1 review rating it a single *.  and it wanted to
know if the review was helpful -- and it wasn't --
as it doesn't say *what* version the review was about.

The reviews seem to be generally "worthless", if they don't
tell me which version is being reviewed.  A comment about a module
from V0.0.1 isn't very useful for a module at version 2.3.7 (as a
randomly chosen example). 


This brings up a few requests/ideas for improvement:

1) Could the version number a review refers to be included?

For comments already posted, either, mark them as too old to
to have included the reviewed version, OR, *IF* the date
of the review is known (it probably should be included in the
review anyway), then match it to the highest version before
the review was added.

2) There should be a way of "closing out" reviews that would have
really been better placed as bugs or RFE's, that have been addressed
(i.e. bug-fixed, or feature-added, etc...).

As it stands now, reviews don't say anything about current versions
and for "bug-fix" reviews, there's no way to indicate the bug was
fixed and the review doesn't apply to a more current version.  There
is also no way to respond to a review to mention that anything has changed
or been fixed.  Any response ends up as another, _unconnected_ comment --
not very useful.

Thoughts?  Ideas/feedback?
Thanks...

[toc] | [next] | [standalone]

#3294

> On Sep 9, 2016, at 3:50 AM, L. A. Walsh <perl-diddler@tlinx.org> wrote:
> 
> Dan Book wrote:
>> 
>> These issues are mostly regarding the cpanratings site itself, which is not part of search.cpan.org <http://search.cpan.org/>. You can find its issue tracker here: https://github.com/perlorg/perlweb/issues <https://github.com/perlorg/perlweb/issues>
> ----
>   Since search.cpan.org includes the ratings as part of its display of
> results, I find it hard to consider them separate.
>> 
>> And you might also find this discussion interesting, regarding similar links from metacpan.org <http://metacpan.org/>: https://github.com/metacpan/metacpan-web/issues/1653 <https://github.com/metacpan/metacpan-web/issues/1653>
> ---
>   Was interesting for a bit -- but it seems like the discussion
> sorta drizzled out w/o any decisions as it appears the stakeholders are
> not willing to participate.

The stakeholders are certainly willing to participate, but to be honest, there are much more pressing issues to be resolved on the MetaCPAN side before we start messing with the display or non-display of reviews.  If there's something resembling consensus and an accompanying appropriate pull request, we'll be happy to deal with it.  The conversation was initially derailed, but it has continued in private.  We're well aware of the problem, but it's just low on the list.


>  Since the cpanratings people aren't willing
> to even participate in fixing the problems, the ratings should, *at least*
> be removed from the search site -- and hopefully the module-info page,
> as they don't represent useful information about the modules (IMO, as they
> don't say what version the comments were about). 

At least one patch has already been applied to CPANRatings as a result of the discussion.  It's not unreasonable to assume that more would be welcome.

Olaf

[toc] | [prev] | [next] | [standalone]

#3297

Olaf Alders wrote:
> The stakeholders are certainly willing to participate, ... 
> The conversation was initially derailed, but it has continued in private.
----
    This could hardly be expected to be known outside of said discussion.
> At least one patch has already been applied to CPANRatings as a result of the discussion.  It's not unreasonable to assume that more would be welcome.
>   
---
    Not clear what the patch did, but how much effort would it be to remove
code referencing 'cpanratings' from cpan until some future discussion 
perfects
the issues discussed?  I.e. it appears that it might have been premature to
release a ratings feature, that appears more like Facebook or Google's 
"Like",
than a technical review/rating of cpan modules.

    If one wants to submit patches, I'm not aware of where the source
to the cpan 'site' is posted such that one could submit diffs against it.
Is it a module in CPAN? :-) 

[toc] | [prev] | [next] | [standalone]

#3299

> On Sep 10, 2016, at 7:01 AM, L. A. Walsh <perl-diddler@tlinx.org> wrote:
> 
> Olaf Alders wrote:
>> The stakeholders are certainly willing to participate, ... The conversation was initially derailed, but it has continued in private.
> ----
>   This could hardly be expected to be known outside of said discussion.

Indeed.  That's why I raised it here.  :)

>> At least one patch has already been applied to CPANRatings as a result of the discussion.  It's not unreasonable to assume that more would be welcome.
>>  
> ---
>   Not clear what the patch did,

https://github.com/perlorg/perlweb/pull/187

It makes it easier to get per-distribution info from the CPANRatings API

> but how much effort would it be to remove
> code referencing 'cpanratings' from cpan until some future discussion perfects
> the issues discussed?  I.e. it appears that it might have been premature to
> release a ratings feature, that appears more like Facebook or Google's "Like",
> than a technical review/rating of cpan modules.

This has been in MetaCPAN since early times, so I'm guessing maybe 5 years that it has been around?  It's a bit late to just back out of the change.  I'd rather get consensus on how to improve it rather than to gut it entirely, unless there's consensus on just removing it.


> 
>   If one wants to submit patches, I'm not aware of where the source
> to the cpan 'site' is posted such that one could submit diffs against it.
> Is it a module in CPAN? :-) 

https://github.com/metacpan/metacpan-developer

Pull requests are very much appreciated, but it's generally a good idea to discuss patches with us first, so that everybody is on the same page.

Olaf

[toc] | [prev] | [next] | [standalone]

#3300

Olaf Alders wrote:
>
> This has been in MetaCPAN since early times, so I'm guessing maybe 5 years that it has been around?  It's a bit late to just back out of the change.  I'd rather get consensus on how to improve it rather than to gut it entirely, unless there's consensus on just removing it.
>   
----
    Well considering it's, _essentially_, "corrupt" data, any use of it 
is likely to lead to incorrect results. 

That doesn't seem like something that should be kept active.  If nothing
else, it will draw attention from those using to the problems and motivate
coming up with a fixed/better solution.

Besides,  It's not like it is a feature in perl; the main request, at this
point,  was to hide cpan.search.org's use of it until the problems are
sorted out. 

    It seems to me, that if any process is relying on that data, it's 
likely suffering from GIGO, which makes me wonder what benefit it could 
be providing
such that it needs to stay active (?)...

[toc] | [prev] | [next] | [standalone]

#3301

> On Sep 10, 2016, at 7:11 PM, L. A. Walsh <perl-diddler@tlinx.org> wrote:
> 
> Olaf Alders wrote:
>> 
>> This has been in MetaCPAN since early times, so I'm guessing maybe 5 years that it has been around?  It's a bit late to just back out of the change.  I'd rather get consensus on how to improve it rather than to gut it entirely, unless there's consensus on just removing it.
>>  
> ----
>   Well considering it's, _essentially_, "corrupt" data, any use of it is likely to lead to incorrect results. 
> That doesn't seem like something that should be kept active.  If nothing
> else, it will draw attention from those using to the problems and motivate
> coming up with a fixed/better solution.
> 
> Besides,  It's not like it is a feature in perl; the main request, at this
> point,  was to hide cpan.search.org's use of it until the problems are
> sorted out. 

If you're looking to patch search.cpan.org, that's an entirely different matter. I don't know what the process is for patching that site.

Olaf

[toc] | [prev] | [next] | [standalone]

#3302

Olaf Alders wrote:
> If you're looking to patch search.cpan.org, that's an entirely different matter. I don't know what the process is for patching that site.
>   
----
Um...Subject?

I thought we were talking about placement in the search
results.  What did you think I was talking about?  meta?  That's a much
larger issue, as the ratings and reviews are a central part of that
site -- and figuring out how meta should function and look is, indeed,
a much more complicated issue...

[toc] | [prev] | [next] | [standalone]

#3307

On Fri, Sep 09, 2016 at 12:50:49AM -0700, L. A. Walsh wrote:

>    Was interesting for a bit -- but it seems like the discussion
> sorta drizzled out w/o any decisions as it appears the stakeholders are
> not willing to participate.  Since the cpanratings people aren't willing
> to even participate in fixing the problems, the ratings should, *at least*
> be removed from the search site ...

I believe that if you want changes made to search.cpan.org you need to
talk to Graham Barr.

-- 
David Cantrell | top google result for "internet beard fetish club"

    Erudite is when you make a classical allusion to a
    feather.  Kinky is when you use the whole chicken.

[toc] | [prev] | [standalone]

Back to top | Article view | perl.module-authors

csiph-web