[Catalog-sig] More problems with the comments system...

Tarek Ziadé ziade.tarek at gmail.com
Wed Nov 25 18:35:34 CET 2009

On Wed, Nov 25, 2009 at 6:04 PM, Martijn Faassen <faassen at startifact.com> wrote:
> Martin v. Löwis wrote:
>>> In this discussion, I placed a comment in the zope.app.sqlscript thread
>>> to point to more modern ways to integrate the Zope Toolkit with
>>> relational databases. This comment (along with many others) is now lost
>>> because the original person deleted his comment and then recreated it.
>> It's not completely lost, as there still is a log of it somewhere. So if
>> you need the actual text, please let me know.
> Thanks, but that isn't a problem.
>>> The current design therefore gives the commenter the power to destroy
>>> responses by package owners. And...
>> That's true, and unfortunate. Can you propose an alternative approach
>> (short of removing the comment facility altogether)?
> Perhaps when someone removes their comment it will say "comment removed"
> (and the commenter's name could be eliminated as well). The followups would
> remain in-tact this way. That isn't the complete context of course for
> future readers, but better than nothing.
>>>> - Because I'm marked as a "Package Index Owner", I can't rate or
>>>> independently comment on this package, even though I'm not its
>>>> maintainer, I'm just someone who's volunteered to take responsibility
>>>> for releasing a load of zope-related eggs when no-one else is around.
>>> Happened to me too.
>> Not sure what exactly happened: that you wanted to rate, and couldn't,
>> or that you wanted to comment, and couldn't?
> I wanted to comment, but couldn't. I can only reply to comments as a package
> owner, not create new comments.
>> Why would you want to comment on the package, when you can put whatever
>> you want to say into the packages' page?
> I wanted to comment on the package itself, because my comment got removed in
> a comment on a comment. I wanted to put this information back in.
> Of course often updating the long description is an alternative to doing so.
> But this would have been quite a bit more work than placing a comment; the
> long description of this package is generated by setup.py of the package,
> and it'd meant having to check out the package and doing a new description
> upload. I just wanted to spend a minute to provide helpful information.
> [snip]
>>> If comment disabling is implemented, I think a nice feature might be to
>>> repeat the author email metadata in its place (or perhaps a special
>>> metadata field for diccussion forums). This way someone who wants to
>>> comment on the package gets a clear indication of where they can to go.
>> This I cannot understand. Can you rephrase?
> If it's possible to disable comments on a per-package basis, it might be
> useful to say in the UI:
> "If you want to comment on this package, please use the following forum:
> <mailing list address or http link>."

Note that I have introduced three new fields in PEP 345 to help on this (after
I have read Catalog-SIG threads on the commenting issues -- someone
suggested them):


# Repository-URL
# Repository-Browser-URL
# Bug-Tracker-URL

Any feedback is welcome about it at Distutils-SIG, if you see anything else
that could be done at the metadata level (then used and highlighted by PyPI)


More information about the Catalog-SIG mailing list