[Catalog-sig] Package comments

["Martin v. Löwis"]

> Comments face a related issue which exposes the psychology a bit more
> clearly: there's a barrier to commenting (you have to register)

In the current implementation on PyPI, you also have to register to
participate in the rating.

> In my experience, this
> results in the opposite of the YouTube phenomenon: comments will tend
> to cluster strongly around the negative extreme of the spectrum
> (people who are very upset about something are generally willing to go
> to more effort to express their feeling than people who are very happy
> about it).

In the case of PyPI, experience so far shows the opposite: comments
are, in general, supportive of the package.

> * Since package maintainers are typically already pressed for time,
>   it's unlikely they'll participate much even if granted the ability
>   to do so, preferring instead the discussion and support channels
>   they've already set up.

But this misses totally the point of the feature. If I have to research
the support channels first, and filter out technical discussions to
actually get to opinions, I might as well evaluate the package myself
directly.

> * But since package maintainers cannot currently respond to comments,

That's not true - they can.

> * Since package maintainers also cannot remove abusive comments

That's not true - they can issue a support request to remove a comment.

> * Since package maintainers have no way -- short of ad-hoc attempts to
>   abuse package descriptions -- to point users toward more useful
>   discussion forums, it's likely that at least some users will never
>   discover such forums.

So you mean, the users have no way of finding out what the
support channels are? How are they then supposed to use them?

Regards,
Martin