[Catalog-sig] Package comments

M.-A. Lemburg mal at egenix.com
Fri Nov 6 10:49:09 CET 2009

Laura Creighton wrote:
> In a message of Thu, 05 Nov 2009 11:32:28 +0100, "M.-A. Lemburg" writes:
> <snip>
>> So you are suggesting to keep the comments and remove the rating
>> system ?
>> Wouldn't that lead to the usual tail of comments you see on
>> blog entries instead of encouraging reviews ?
>> -- 
>> Marc-Andre Lemburg
>> eGenix.com
> I think that a rating system is a bad idea, period.  I think that
> the comments feature _could_ be a good idea, if used well, though I
> think that the design should be reworked if what we want to have
> is a way to centrally find long reviews of software packages.
> The thing is, I now believe that this is not Martin van Löwis goal.
> I think that he wants to make things easier for software downloaders
> to evaluate software, no matter how unpopular these ideas are with
> the package creators. 

While I can't speak for Martin, I don't think he has had any
such intentions. The feature was requested on the PyPI tracker
and he implemented it. That's all.

> I think that this is a bad decision; it is
> the software creators one should be most concerned with.  I don't
> think that a rating system actually serves the downloaders, and to
> the extent that it drives away package creators, it actively harms
> them.

There will certainly be some developers that don't like being
rated for things they put on PyPI as a service to the community.

Others will probably take it as challenge to get a high rating.

Yet others will just not really care.

And then you have people who download and install packages
via e.g. pip or easy_install and don't get to see all this
in the first place.

> There are things we could do -- add more tags for 'how easy is this
> to install' and 'where do you want discussion of this package to
> occur' and 'how well maintained is this package'.  But I think there
> is a great need for packages that were written as a one-shot for
> somebody who needed to solve some particular problem, aren't being
> maintained, and are very much 'I am releasing this in case somebody
> else finds it useful'.  I often find stuff like this useful, and I
> fear my supply of such things will dry up if people have to face
> comments and criticism that they do not want.

I'm not sure. Most of these discussions sound like FUD to me.

The only point I'd like to make is that PyPI should not force any
such decision on the developers with no way for them to opt-in or

IMHO, PyPI should offer comments and rating as opt-in
possibility for developers who want to use these features.

Unless it's possible to reach out to all PyPI developers via email to
inform them of enabling the feature by default, it should be disabled
per default - it's unfair to enable a controversial feature
without their consent.

Regarding the rating system itself, my only request is to publicly
show the per-user comment ratings (they are currently hidden) in
order to get more transparency. For uncommented ratings, an empty
entry would be needed in order to achieve the same level of

Marc-Andre Lemburg

Professional Python Services directly from the Source  (#1, Nov 06 2009)
>>> Python/Zope Consulting and Support ...        http://www.egenix.com/
>>> mxODBC.Zope.Database.Adapter ...             http://zope.egenix.com/
>>> mxODBC, mxDateTime, mxTextTools ...        http://python.egenix.com/

::: Try our new mxODBC.Connect Python Database Interface for free ! ::::

   eGenix.com Software, Skills and Services GmbH  Pastor-Loeh-Str.48
    D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg
           Registered at Amtsgericht Duesseldorf: HRB 46611

More information about the Catalog-SIG mailing list