[Catalog-sig] Please turn off ratings
fuzzyman at gmail.com
Wed Apr 6 13:55:45 CEST 2011
On 6 April 2011 08:21, James Bennett <ubernostrum at gmail.com> wrote:
> On Wed, Apr 6, 2011 at 1:13 AM, Terry Reedy <tjreedy at udel.edu> wrote:
> > But it seems to me likely biased, so I am not convinced yet.
> There is not and never will be a sampling of "end users" that will be
> widely accepted by this list. This is a fact that we need to accept
> and move past, largely because such sampling is a red herring.
> > An avalanche follows. It mostly consists of people who did not ratings
> > before saying they do not like ratings now, and perhaps even less.
> > And there is some rehashing of the same old arguments.
> When I look at this and the original threads, I see a variety of
> opinions. There are people (I'm one of them) who feel PyPI should not
> be in the business of hosting a rating system, and should simply stick
> to being the Python Package Index. There are people who feel there
> could be a useful rating implementation, but that the one PyPI has is
> not such an implementation. And there's Martin.
I think this is unfair and personally antagonistic. I'm surprised (and a
little ashamed) at the level of emotion and invective that this thread has
I don't see what harm the rating system does and it has been demonstrated in
this thread that they *can* be useful.
When rating and comments were introduced it was made clear they were not for
package authors but for PyPI users. Guido himself weighed in on the
discussion and said that he thought comments and ratings were useful
features for users. No they're not useful for package authors as a feedback
mechanism, but that isn't their intent.
I don't find the ratings system particularly useful as a package author or
as a pypi user, but I have no problem with it existing.
> There seems to be virtually universal agreement that -- if PyPI is to
> host ratings -- the current implementation is flawed at best. This is
> a strong argument for scrapping the implementation and doing the
> legwork to get something that's actually demonstrably helpful. It is
> not an argument for keeping the status quo.
"Not perfect" is not an argument for scrapping something without an
alternative. It seems no-one in this thread is willing to discuss or
countenance alternatives, until that happens calling for Martin's head on a
platter is just not helpful or reasonable. I think Martin's responses in
this thread have been stoic and reasonable.
A possible alternative: allowing users to "star" or "like" projects they
find particularly useful. Either with or without their name being attached
to their response. ("Jacon Kaplan-Moss and 47 other users liked this
project" for example.)
For reference, some of Guido's thoughts on the topics:
All the best,
> But that, ultimately, is what the problem is here. What we're seeing
> is not a debate or discussion. What we're seeing is one person with
> authority -- Martin -- stonewalling the rest of the community in an
> attempt to preserve the status quo. No amount of polling of users will
> establish the utility or desirability of ratings; we've seen ample
> evidence of that in the fact that A) allegedly ratings are popular
> enough to win a poll, but B) ratings are so unpopular that practically
> nobody actually submits ratings. Suggestions of further polling are
> simply a stalling tactic, dragging this out long enough that the
> people who don't like ratings will simply give up from frustration or
> exhaustion, at which point things stay as they are.
> Add to that the fact that, as I've repeatedly pointed out, package
> maintainers -- the people who make PyPI worthwhile -- are essentially
> being told that their concerns will be ignored, and that any arguments
> package maintainers make against ratings (or, in the beginning,
> comments -- remember, when we asked for the ability to toggle them we
> were equated to government censors) were simply shouted down as a
> failure on *our* part to compromise, and, well, this isn't a pretty
> This is a horrifically dysfunctional way to manage a valuable
> community resource. It needs to stop. It needs to stop now. I don't
> care how bad a taste that ends up leaving in someone's mouth, or how
> many emails René has to passive-aggressively complain about receiving.
> I care about the future of PyPI and the communities -- BOTH developers
> and end users -- it serves, and right now that future looks awfully
> "Bureaucrat Conrad, you are technically correct -- the best kind of
> Catalog-SIG mailing list
> Catalog-SIG at python.org
May you do good and not evil
May you find forgiveness for yourself and forgive others
May you share freely, never taking more than you give.
-- the sqlite blessing http://www.sqlite.org/different.html
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Catalog-SIG