[Catalog-sig] The removal of ratings sucks

Tres Seaver tseaver at palladion.com
Thu Apr 28 15:16:30 CEST 2011

Hash: SHA1

On 04/28/2011 07:06 AM, Marcello Perathoner wrote:
> I'm upset, because I *was* using the rating system. I also *knew* the 
> limitations of this feature. I didn't think of choosing an application 
> framework based on the rating. I'm not that stupid. (But I have just 
> learned that some people believe that I'm that stupid. Thank you!)
> The rating system helped me choose the best candidate for quick hacks, 
> but sadly this "battery" is not included anymore.
> When I'm adding a simple feature that should not take more than 5 
> minutes to do, I don't want to compare package features, nor see the 
> package's web site, nor see how many tickets are open, nor if the 
> maintainer answers their email ... I want what other users think is the 
> overall best package.

How did you evaluate the care or thought put into the ratings by the
tiny handful of PyPI users who actually created them?  "Crowd-sourcing"
relies on high participation to correct for outliers:  without lots of
ratings, any one bogus rating, whether derogatory or flattering, carries
too much weight.

> I'm surprised that this useful feature has been removed.
> I'm even more surprised, after reading the relative thread, that the 
> feature has been removed in an effort to kill the messenger by package 
> maintainers that got bad ratings. How could that happen after there was 
> a (pretty fair) poll that showed the majority was wanting a feature like 
> that?

I don't agree with that characterization of the thread at all:  the
argument was not "my package got a bad rating, remove ratings", but
"quasi-anonymous ratings don't supply enough value".  Jacob, for
instance, had mostly very high ratings for Django, but had no
information on *why* a particular bad rating occurred, and could
therefore not address the problem.

> Is PyPI run by a vociferous minority?

PyPI is run by MvL on behalf of the Python community, with support and
feedback (and often not enough gratitude) from the members of this list.
 In the end, I believe MvL was convinced to remove ratings not by
arguments about their usefulness, but by the evidence that almost nobody
rated packages.

- -- 
Tres Seaver          +1 540-429-0999          tseaver at palladion.com
Palladion Software   "Excellence by Design"    http://palladion.com
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/


More information about the Catalog-SIG mailing list