On Tue, Apr 5, 2011 at 4:07 PM, "Martin v. Löwis" <mar...@v.loewis.de> wrote: > Because the users of the service have actually requested that it works > this way.
I've not seen this. Citations, please? You just reminded me, quite correctly, that I'm not exactly representative of the average PyPI user, and you claimed that this makes me unqualified to speak on their behalf. And yet you seem perfectly comfortable speaking on behalf of the users. Do you have some secret feedback channel I could get some access to? Just as a quick point of reference, I asked on Twitter: "PyPI allows users to rate packages. Does anyone use this feature? Do you find it useful? Why?" (http://twitter.com/jacobian/status/55370416444809216) So far, the answers have been: "@jacobian I almost never actually visit PyPI, my interactions with it are almost entirely moderated by pip. So, not that useful to me..." (http://twitter.com/joshourisman/status/55370598406299649) "@jacobian My thoughts on that are in the third paragraph of http://pydanny.blogspot.com/2011/04/pycon-2011-sprint-report.html"; (http://twitter.com/pydanny/status/55371268664463361) "@jacobian Does anyone use it: not enough to be meaningful. Do I find it useful: no. Why: What criteria should ratings even follow?" (http://twitter.com/schmichael/status/55371725076045825) "@jacobian PyPI ratings are stupid and not useful. The PyPI ratings rate compare packages in relation to what, exactly? That's the problem." (http://twitter.com/howardbutler/status/55372373351862272) "@jacobian Too hidden and nowhere used on the site in terms of top lists or sorting." (http://twitter.com/passy/status/55373293749940224) "@jacobian No, I don't use it. If anything counts than it's the number of downloads." (http://twitter.com/keimlink/status/55374431136120833) "@jacobian I didn't know the feature existed, but I could see it being really useful if it were more exposed or even available from pip." (http://twitter.com/defrex/status/55373820718100480) "@jacobian no, I don't use it, I ignore the ratings and comments." (http://twitter.com/jessenoller/status/55375748403445760) "@jacobian I voiced my strong dislike of the feature at the time on catalog-sig. I don't use it, whether as user or as package author." (http://twitter.com/jezdez/status/55377232327213056) "@jacobian Never really got the point of rating packages. Optimally they all solve different problems and can't be compared anyway" (http://twitter.com/DasIch/status/55377740005773313) To my eyes, that's 9 -1s and one +0/-0. Granted, my Twitter followers are self-selecting, but I do have quite a few of them and if the feature was as fastly popular as you say wouldn't at least one person who follows me speak up? I also asked on a variety of IRC channels I hang out on. Most are semi-private so I'm not going to quote anyone here, but again not a single person has given anything other than a "meh", and most seem distinctly unhappy about the feature. I don't think the support for this feature you claim exists actually does. > PyPI has grown features over time because of user requests, I have no problem with that. Now users are requesting that you remove a broken feature. > such as hosting files (which it originally didn't do - it was a mere > catalog, not a repository), and (more recently) hosting documentation. > Do you think file hosting turns it into something political? Do you really not see the difference between file hosting -- something that, if I don't need I don't have to use -- and a forced-upon-authors-by-fiat rating system? Jacob _______________________________________________ Catalog-SIG mailing list Catalog-SIG@python.org http://mail.python.org/mailman/listinfo/catalog-sig
