From techtonik at gmail.com Sat Apr 2 21:07:35 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Sat, 2 Apr 2011 22:07:35 +0300 Subject: [Catalog-sig] last.pypi.python.org Message-ID: http://pypi.python.org/mirrors/ refers to last.pypi.python.org to query for a last mirror, but it leads to a broken page. What is the status of it? -- anatoly t. From merwok at netwok.org Sun Apr 3 18:07:41 2011 From: merwok at netwok.org (=?UTF-8?B?w4lyaWMgQXJhdWpv?=) Date: Sun, 03 Apr 2011 18:07:41 +0200 Subject: [Catalog-sig] last.pypi.python.org In-Reply-To: References: Message-ID: <4D989B4D.2060508@netwok.org> Le 02/04/2011 21:07, anatoly techtonik a ?crit : > http://pypi.python.org/mirrors/ refers to last.pypi.python.org to > query for a last mirror, but it leads to a broken page. This domain name is supposed to be used for CNAME resolution, not HTTP requests. See PEP 381. Regards From martin at v.loewis.de Sun Apr 3 18:50:14 2011 From: martin at v.loewis.de (=?UTF-8?B?Ik1hcnRpbiB2LiBMw7Z3aXMi?=) Date: Sun, 03 Apr 2011 18:50:14 +0200 Subject: [Catalog-sig] last.pypi.python.org In-Reply-To: <4D989B4D.2060508@netwok.org> References: <4D989B4D.2060508@netwok.org> Message-ID: <4D98A546.3020305@v.loewis.de> Am 03.04.2011 18:07, schrieb ?ric Araujo: > Le 02/04/2011 21:07, anatoly techtonik a ?crit : >> http://pypi.python.org/mirrors/ refers to last.pypi.python.org to >> query for a last mirror, but it leads to a broken page. > > This domain name is supposed to be used for CNAME resolution, not HTTP > requests. Still, it's easy to fix to make it work for direct http requests also, so I just did. Regards, Martin From tjreedy at udel.edu Mon Apr 4 21:32:11 2011 From: tjreedy at udel.edu (Terry Reedy) Date: Mon, 04 Apr 2011 15:32:11 -0400 Subject: [Catalog-sig] last.pypi.python.org In-Reply-To: <4D98A546.3020305@v.loewis.de> References: <4D989B4D.2060508@netwok.org> <4D98A546.3020305@v.loewis.de> Message-ID: On 4/3/2011 12:50 PM, "Martin v. L?wis" wrote: > Am 03.04.2011 18:07, schrieb ?ric Araujo: >> Le 02/04/2011 21:07, anatoly techtonik a ?crit : >>> http://pypi.python.org/mirrors/ refers to last.pypi.python.org to >>> query for a last mirror, but it leads to a broken page. >> >> This domain name is supposed to be used for CNAME resolution, not HTTP >> requests. > > Still, it's easy to fix to make it work for direct http requests also, > so I just did. Not knowing what CNAME means, I might have naively made the same mistake of trying out last.pypi.python.org, which does indeed work now. -- Terry Jan Reedy From martin at v.loewis.de Mon Apr 4 22:49:42 2011 From: martin at v.loewis.de (=?UTF-8?B?Ik1hcnRpbiB2LiBMw7Z3aXMi?=) Date: Mon, 04 Apr 2011 22:49:42 +0200 Subject: [Catalog-sig] last.pypi.python.org In-Reply-To: References: <4D989B4D.2060508@netwok.org> <4D98A546.3020305@v.loewis.de> Message-ID: <4D9A2EE6.7080008@v.loewis.de> Am 04.04.2011 21:32, schrieb Terry Reedy: > On 4/3/2011 12:50 PM, "Martin v. L?wis" wrote: >> Am 03.04.2011 18:07, schrieb ?ric Araujo: >>> Le 02/04/2011 21:07, anatoly techtonik a ?crit : >>>> http://pypi.python.org/mirrors/ refers to last.pypi.python.org to >>>> query for a last mirror, but it leads to a broken page. >>> >>> This domain name is supposed to be used for CNAME resolution, not HTTP >>> requests. >> >> Still, it's easy to fix to make it work for direct http requests also, >> so I just did. > > Not knowing what CNAME means, I might have naively made the same mistake > of trying out last.pypi.python.org, which does indeed work now. It's not a problem to be naive here: If you don't know what a CNAME is, you weren't target audience of last.pypi.python.org. Hence it's not a problem that you get a bogus or no page when you put that into a browser. I promise that such breakage will happen again, and don't plan to make fixing it a priority. Regards, Martin From monitor at jacobian.org Tue Apr 5 09:30:01 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Tue, 05 Apr 2011 02:30:01 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection failed Message-ID: <1301988606.1@jacobian.org> Connection failed Service pypi.python.org Date: Tue, 05 Apr 2011 02:30:01 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Tue Apr 5 10:05:39 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Tue, 05 Apr 2011 03:05:39 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection succeeded Message-ID: <1301990743.1@jacobian.org> Connection succeeded Service pypi.python.org Date: Tue, 05 Apr 2011 03:05:39 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[pypi.python.org:80] via TCP Your faithful employee, monit From jacob at jacobian.org Tue Apr 5 20:28:26 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 13:28:26 -0500 Subject: [Catalog-sig] Please turn off ratings Message-ID: Hi folks -- Just got another one of these: > From: PyPI operators > Subject: New rating on Django > > [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] > xen has rated your package as 0/5. ?Comment (optional): I'd like to ask, again, to turn off ratings on PyPI packages. How is this feature helpful to anyone at all? * It's not helpful to me as a maintainer: I have no idea *why* "xen" gave this rating. * It's not helpful to users: they have no idea what "0" means. * It's not helpful to "xen": his feedback can't be acted upon, so we'll never be able to aprove. * It's not helpful to PyPI: what value does "0/5" provide to a catalog? * It's not helpful to the Python language: how does "0/5" in any way help Python users choose tools? PyPI is a catalog. Naive rating/voting features aren't part of that mission. Can we please turn this ill-considered, useless feature off? Jacob From noah at coderanger.net Tue Apr 5 20:37:00 2011 From: noah at coderanger.net (Noah Kantrowitz) Date: Tue, 5 Apr 2011 11:37:00 -0700 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: <034a01cbf3c0$73f9eb20$5bedc160$@net> > -----Original Message----- > From: catalog-sig-bounces+noah=coderanger.net at python.org > [mailto:catalog-sig-bounces+noah=coderanger.net at python.org] On Behalf > Of Jacob Kaplan-Moss > Sent: Tuesday, April 05, 2011 11:28 AM > To: catalog-sig > Subject: [Catalog-sig] Please turn off ratings > > Hi folks -- > > Just got another one of these: > > > From: PyPI operators > > Subject: New rating on Django > > > > [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] > > xen has rated your package as 0/5. ?Comment (optional): > > I'd like to ask, again, to turn off ratings on PyPI packages. > > How is this feature helpful to anyone at all? > > * It's not helpful to me as a maintainer: I have no idea *why* "xen" > gave this rating. > * It's not helpful to users: they have no idea what "0" means. > * It's not helpful to "xen": his feedback can't be acted upon, so > we'll never be able to aprove. > * It's not helpful to PyPI: what value does "0/5" provide to a catalog? > * It's not helpful to the Python language: how does "0/5" in any way > help Python users choose tools? > > PyPI is a catalog. Naive rating/voting features aren't part of that > mission. Can we please turn this ill-considered, useless feature off? +1 --Noah From jim at zope.com Tue Apr 5 20:44:35 2011 From: jim at zope.com (Jim Fulton) Date: Tue, 5 Apr 2011 14:44:35 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: On Tue, Apr 5, 2011 at 2:28 PM, Jacob Kaplan-Moss wrote: > Hi folks -- > > Just got another one of these: > >> From: PyPI operators >> Subject: New rating on Django >> >> [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] >> xen has rated your package as 0/5. ?Comment (optional): > > I'd like to ask, again, to turn off ratings on PyPI packages. > > How is this feature helpful to anyone at all? > > * It's not helpful to me as a maintainer: I have no idea *why* "xen" > gave this rating. > * It's not helpful to users: they have no idea what "0" means. > * It's not helpful to "xen": his feedback can't be acted upon, so > we'll never be able to aprove. > * It's not helpful to PyPI: what value does "0/5" provide to a catalog? > * It's not helpful to the Python language: how does "0/5" in any way > help Python users choose tools? > > PyPI is a catalog. Naive rating/voting features aren't part of that > mission. Can we please turn this ill-considered, useless feature off? +1 Jim -- Jim Fulton http://www.linkedin.com/in/jimfulton From kevin at bud.ca Tue Apr 5 20:49:02 2011 From: kevin at bud.ca (Kevin Teague) Date: Tue, 5 Apr 2011 11:49:02 -0700 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: Django 1.3 has a score of 0/5 but Django 1.2.5 has a score of 5/5. Either +1 or Django has really been dropping the ball in the QA dept. -------------- next part -------------- An HTML attachment was scrubbed... URL: From jacob at jacobian.org Tue Apr 5 20:58:34 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 13:58:34 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: On Tue, Apr 5, 2011 at 1:49 PM, Kevin Teague wrote: > Django 1.3 has a score of 0/5 but Django 1.2.5 has a score of 5/5. > Either +1 or Django has really been dropping the ball in the QA dept. Heh. I should point out, in case it's not obvious, that it's not really the 0/5 that I'm objecting to. Yeah, that stings, but 5/5 isn't any more useful. Again, it doesn't provide any real feedback, and so while 5/5 is nicer to my ego it's still useless as a form of feedback. Jacob From bkjones at gmail.com Tue Apr 5 21:49:52 2011 From: bkjones at gmail.com (Brian Jones) Date: Tue, 5 Apr 2011 15:49:52 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: On Tue, Apr 5, 2011 at 2:58 PM, Jacob Kaplan-Moss wrote: > Again, it doesn't provide any real feedback, and so while 5/5 > is nicer to my ego it's still useless as a form of feedback. > +1 - well stated. > > Jacob > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- Brian K. Jones My Blog http://www.protocolostomy.com Follow me http://twitter.com/bkjones -------------- next part -------------- An HTML attachment was scrubbed... URL: From martin at v.loewis.de Tue Apr 5 21:55:29 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 21:55:29 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: <4D9B73B1.2050804@v.loewis.de> > How is this feature helpful to anyone at all? I maintain that the answer to that question is the same as before: it is helpful to the users of the package. I can easily see that it is not that helpful to package authors (in particular when they turned off comments). In any case, the community had agreed to to allow rating but to make commenting optional (at the package author's choice). I don't see what has changed since. Regards, Martin From lac at openend.se Tue Apr 5 22:32:13 2011 From: lac at openend.se (Laura Creighton) Date: Tue, 05 Apr 2011 22:32:13 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: Message from =?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?= of "Tue, 05 Apr 2011 21:55:29 +0200." <4D9B73B1.2050804@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> Message-ID: <201104052032.p35KWDA5019589@theraft.openend.se> In a message of Tue, 05 Apr 2011 21:55:29 +0200, "Martin v. L?wis" writes: >> How is this feature helpful to anyone at all? > >I maintain that the answer to that question is the same as before: >it is helpful to the users of the package. I can easily see that >it is not that helpful to package authors (in particular when they >turned off comments). > >In any case, the community had agreed to to allow rating but to >make commenting optional (at the package author's choice). I don't >see what has changed since. > >Regards, >Martin I said it was a bad idea at the time. Now that more people have had a chance to see it in action, more of them agree with me. That's what changed. Laura From martin at v.loewis.de Tue Apr 5 22:36:34 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 22:36:34 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <201104052032.p35KWDA5019589@theraft.openend.se> References: <4D9B73B1.2050804@v.loewis.de> <201104052032.p35KWDA5019589@theraft.openend.se> Message-ID: <4D9B7D52.3090606@v.loewis.de> > I said it was a bad idea at the time. Now that more people have had a > chance to see it in action, more of them agree with me. That's what > changed. So how can we find out what the PyPI users want? Regards, Martin From jacob at jacobian.org Tue Apr 5 22:38:55 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 15:38:55 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B73B1.2050804@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 2:55 PM, "Martin v. L?wis" wrote: > I maintain that the answer to that question is the same as before: > it is helpful to the users of the package. How? Am I dense? I fail to see what "0/5" or "3/5" or "5/5" does. It's a piece of data, sure, but so is 17, 10045.223, and 4/7. But none of these provide any *information*. > In any case, the community had agreed to to allow rating but to > make commenting optional (at the package author's choice). I don't > see what has changed since. Who is this "community" you speak of, and when and where did "they" agree? I don't see a single person other than you stepping up to defend this ill-considered feature. Jacob From martin at v.loewis.de Tue Apr 5 22:42:31 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 22:42:31 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> Message-ID: <4D9B7EB7.3070207@v.loewis.de> >> In any case, the community had agreed to to allow rating but to >> make commenting optional (at the package author's choice). I don't >> see what has changed since. > > Who is this "community" you speak of, and when and where did "they" > agree? We performed a user poll, which came with the status quo. Regards, Martin From jacob at jacobian.org Tue Apr 5 22:43:02 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 15:43:02 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B7D52.3090606@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <201104052032.p35KWDA5019589@theraft.openend.se> <4D9B7D52.3090606@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 3:36 PM, "Martin v. L?wis" wrote: > So how can we find out what the PyPI users want? You're hearing from them right now. I have 11 packages listed on PyPI. Noah and Jim probably have 2-3x as many each. I'm fairly sure Laura, Brian, and Kevin do as well. We all use PyPI daily. Jacob From jacob at jacobian.org Tue Apr 5 22:44:45 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 15:44:45 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B7EB7.3070207@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> Message-ID: OK, I can tell this is going nowhere; it's turning into me vs. Martin and I don't want that. Martin: what's the next step here? Should I talk to Python-Dev? the PSF? Who has the authority to arbitrate this neutrally and make a decision? Or are you the BDFL for PyPI and is this the final word? Jacob From noah at coderanger.net Tue Apr 5 22:45:30 2011 From: noah at coderanger.net (Noah Kantrowitz) Date: Tue, 5 Apr 2011 13:45:30 -0700 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B7EB7.3070207@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> Message-ID: <038d01cbf3d2$67b77870$37266950$@net> > -----Original Message----- > From: catalog-sig-bounces+noah=coderanger.net at python.org > [mailto:catalog-sig-bounces+noah=coderanger.net at python.org] On Behalf > Of "Martin v. L?wis" > Sent: Tuesday, April 05, 2011 1:43 PM > To: Jacob Kaplan-Moss > Cc: catalog-sig > Subject: Re: [Catalog-sig] Please turn off ratings > > >> In any case, the community had agreed to to allow rating but to > >> make commenting optional (at the package author's choice). I don't > >> see what has changed since. > > > > Who is this "community" you speak of, and when and where did "they" > > agree? > > We performed a user poll, which came with the status quo. Being able to provide feedback to package authors is a useful feature, but this is the wrong implementation (which IMO would explain the user poll results). If you really want something, look at Django Packages' "I use this" counters. That isn't a grade per se, but it provides some level of feedback about the relative merits of similar packages. --Noah From martin at v.loewis.de Tue Apr 5 22:45:54 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 22:45:54 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <201104052032.p35KWDA5019589@theraft.openend.se> <4D9B7D52.3090606@v.loewis.de> Message-ID: <4D9B7F82.80108@v.loewis.de> Am 05.04.2011 22:43, schrieb Jacob Kaplan-Moss: > On Tue, Apr 5, 2011 at 3:36 PM, "Martin v. L?wis" wrote: >> So how can we find out what the PyPI users want? > > You're hearing from them right now. I have 11 packages listed on PyPI. > Noah and Jim probably have 2-3x as many each. I'm fairly sure Laura, > Brian, and Kevin do as well. We all use PyPI daily. Please understand that you still represent a minority of PyPI users only: namely, the package authors. The (by numbers) larger group of users are the ones that actually *use* the packages on PyPI. Of course, trere is also overlap between these groups. Unfortunately, they typically don't subscribe to catalog-sig, so they won't speak up here. Regards, Martin From jim at zope.com Tue Apr 5 22:45:55 2011 From: jim at zope.com (Jim Fulton) Date: Tue, 5 Apr 2011 16:45:55 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B73B1.2050804@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 3:55 PM, "Martin v. L?wis" wrote: >> How is this feature helpful to anyone at all? > > I maintain that the answer to that question is the same as before: > it is helpful to the users of the package. That's fair, but I don't think experience has supported it. I don't think ratings are helpful to Django users. Do you know of any packages for which you think the ratings we have now actually help users? Jim -- Jim Fulton http://www.linkedin.com/in/jimfulton From jacob at jacobian.org Tue Apr 5 22:50:32 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 15:50:32 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B7F82.80108@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <201104052032.p35KWDA5019589@theraft.openend.se> <4D9B7D52.3090606@v.loewis.de> <4D9B7F82.80108@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 3:45 PM, "Martin v. L?wis" wrote: > Please understand that you still represent a minority of PyPI users > only: namely, the package authors. The (by numbers) larger group of > users are the ones that actually *use* the packages on PyPI. Of course, > trere is also overlap between these groups. > > Unfortunately, they typically don't subscribe to catalog-sig, so they > won't speak up here. How is this any different from any other open source project? The majority of Python users don't subscribe to python-dev, yet the core Python team makes decisions on their behalf every day. As maintainers we rely on our ability to "channel" the needs of our community and act according. Jacob From martin at v.loewis.de Tue Apr 5 22:53:14 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 22:53:14 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> Message-ID: <4D9B813A.7010103@v.loewis.de> > Martin: what's the next step here? Should I talk to Python-Dev? the > PSF? Who has the authority to arbitrate this neutrally and make a > decision? Or are you the BDFL for PyPI and is this the final word? It's not the final word. If you can suggest a procedure that fairly involves the end users also (which *frequently* had requested from me that I provide this very feature), and they now say they don't like this, I'd be willing to reconsider. Also, I'd be willing to reconsider if this service was provided offsite, and you could agree that I link to this offsite service (as well as to other services that collect information and opinion about PyPI packages). I'd be also willing to accept an order from the PSF board, but would rather hope that they refrain from giving such an order (except for legal issues). I don't think python-dev has to do anything with this. Regards, Martin From renesd at gmail.com Tue Apr 5 22:56:49 2011 From: renesd at gmail.com (=?ISO-8859-1?Q?Ren=E9_Dudfield?=) Date: Tue, 5 Apr 2011 21:56:49 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: Hi, a small note for anyone who wants to join in painting this shed... Please first see previous recent (long) thread where this was discussed, and where it was decided to leave them on. -1 (out of 5) for turning them off. -------------- next part -------------- An HTML attachment was scrubbed... URL: From martin at v.loewis.de Tue Apr 5 22:59:55 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 22:59:55 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <038d01cbf3d2$67b77870$37266950$@net> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <038d01cbf3d2$67b77870$37266950$@net> Message-ID: <4D9B82CB.70804@v.loewis.de> > Being able to provide feedback to package authors is a useful feature, but > this is the wrong implementation (which IMO would explain the user poll > results). If you really want something, look at Django Packages' "I use > this" counters. That isn't a grade per se, but it provides some level of > feedback about the relative merits of similar packages. Changing the implementation would be fine with me; I modeled this after many other rating systems that had been around back then. However, ISTM that the package authors asking for a change actually accept only wholesale removal of the feature. Regards, Martin From martin at v.loewis.de Tue Apr 5 23:02:03 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:02:03 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> Message-ID: <4D9B834B.4070109@v.loewis.de> > I don't think ratings are helpful to Django users. Do you know of any > packages for which you think the ratings we have now actually help > users? I'd have to search through the package list - but yes, I do think I'd find examples. More likely, packages where comments aren't disabled, since it would allow users to explain how they arrived at their rating. Of course, if there is only a single rating, it's likely not helpful. Regards, Martin From jacob at jacobian.org Tue Apr 5 23:02:36 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 16:02:36 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B813A.7010103@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 3:53 PM, "Martin v. L?wis" wrote: > It's not the final word. If you can suggest a procedure that fairly > involves the end users also (which *frequently* had requested from > me that I provide this very feature), and they now say they don't > like this, I'd be willing to reconsider. Well, there a a variety of ways that users of Django can provide feedback in a constructive way: they can file a ticket, ask for help on django-users, ask questions on Stack Overflow, propose features on django-dev, complain to the DSF, complain to me personally, ... So something simple would be to let package authors provide a "feedback" link. But once again: why is it your position to force package authors to do *anything* here? If I want to put some code up for Python users to use why do I have to follow some arbitrary rules imposed by fiat? I've said it over and over again: PyPI is a catalog. It's a list of Python packages. Anything more dilutes the purpose and turns it into something political. Jacob From jacob at jacobian.org Tue Apr 5 23:05:34 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 16:05:34 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: On Tue, Apr 5, 2011 at 3:56 PM, Ren? Dudfield wrote: > a small note for anyone who wants to join in painting this shed... Please > first see previous recent (long) thread where this was discussed, and where > it was decided to leave them on. > > -1 (out of 5) for turning them off. Perhaps I'm missing something, but I can't find a single articulation of why this feature is actually useful. Can you explain to me: * What, as a package maintainer, I'm supposed to do in response to a 0/5, or 3/5, or 5/5 rating? * What, as a user, I'm supposed to do if I see that Django 1.3 scores 0/5, or 3/5, or 5/5? * What, as a user, I'm supposed to use as a criteria to rate a package? Jacob From ubernostrum at gmail.com Tue Apr 5 23:12:21 2011 From: ubernostrum at gmail.com (James Bennett) Date: Tue, 5 Apr 2011 16:12:21 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B73B1.2050804@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 2:55 PM, "Martin v. L?wis" wrote: >> How is this feature helpful to anyone at all? > > I maintain that the answer to that question is the same as before: > it is helpful to the users of the package. I can easily see that > it is not that helpful to package authors (in particular when they > turned off comments). Notably, you've never answered this question. When asked, you simply assert "it is helpful", without citing a rationale, without providing use cases and without pointing to any evidence of anyone anywhere in the real world actually benefiting from this feature. The onus is on you to demonstrate the benefits. If you can't, the feature should be scrapped. -- "Bureaucrat Conrad, you are technically correct -- the best kind of correct." From martin at v.loewis.de Tue Apr 5 23:14:01 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:14:01 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: <4D9B8619.20802@v.loewis.de> Am 05.04.2011 23:05, schrieb Jacob Kaplan-Moss: > On Tue, Apr 5, 2011 at 3:56 PM, Ren? Dudfield wrote: >> a small note for anyone who wants to join in painting this shed... Please >> first see previous recent (long) thread where this was discussed, and where >> it was decided to leave them on. >> >> -1 (out of 5) for turning them off. > > Perhaps I'm missing something, but I can't find a single articulation > of why this feature is actually useful. Can you explain to me: > > * What, as a package maintainer, I'm supposed to do in response to a > 0/5, or 3/5, or 5/5 rating? It's not primarily targeted at package maintainers - if users like your package, you shouldn't do anything, and if they don't like it, you may not be able to do anything. If users dislike your package, hopefully some would also comment on why they dislike it, in which case you might consider changing your package. > * What, as a user, I'm supposed to do if I see that Django 1.3 scores > 0/5, or 3/5, or 5/5? If there is a single vote only: ignore it. If there are many votes, all voting the package very low, try to find a different package for the same purpose. If there are comments, consider whether they apply to your application as well. > * What, as a user, I'm supposed to use as a criteria to rate a package? If you like the package a lot, give it 5. If you don't like it at all, give it 0. If you are in the middle between these, give some rating in the middle. Is it really that difficult to understand a grade system? In case you are familiar with U.S. grading: A=5, F=0 (if I understand U.S. grading correctly). Regards, Martin From martin at v.loewis.de Tue Apr 5 23:07:53 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:07:53 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> Message-ID: <4D9B84A9.2090904@v.loewis.de> > But once again: why is it your position to force package authors to do > *anything* here? If I want to put some code up for Python users to use > why do I have to follow some arbitrary rules imposed by fiat? > > I've said it over and over again: PyPI is a catalog. It's a list of > Python packages. Anything more dilutes the purpose and turns it into > something political. Because the users of the service have actually requested that it works this way. PyPI has grown features over time because of user requests, 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? Regards, Martin From martin at v.loewis.de Tue Apr 5 23:18:40 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:18:40 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> Message-ID: <4D9B8730.5010106@v.loewis.de> > Notably, you've never answered this question. When asked, you simply > assert "it is helpful", without citing a rationale, without providing > use cases and without pointing to any evidence of anyone anywhere in > the real world actually benefiting from this feature. > > The onus is on you to demonstrate the benefits. If you can't, the > feature should be scrapped. Would it be convincing if users assert that they find it helpful? In abstract, the rationale is that it allows users to select a package without going through the possibly-time-consuming process of evaluating. You want to use software where you know there are many other users who like it. There is a lot of packages on PyPI that is in a rough state, and you don't want to try five packages that all don't work well before settling on the one that everybody else uses. Regards, Martin From jacob at jacobian.org Tue Apr 5 23:19:44 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 16:19:44 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B84A9.2090904@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 4:07 PM, "Martin v. L?wis" 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 From pje at telecommunity.com Tue Apr 5 23:27:13 2011 From: pje at telecommunity.com (P.J. Eby) Date: Tue, 05 Apr 2011 17:27:13 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: <20110405212727.EB71F3A4063@sparrow.telecommunity.com> At 01:28 PM 4/5/2011 -0500, Jacob Kaplan-Moss wrote: >PyPI is a catalog. Naive rating/voting features aren't part of that >mission. Can we please turn this ill-considered, useless feature off? +1 From martin at v.loewis.de Tue Apr 5 23:28:05 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:28:05 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> Message-ID: <4D9B8965.1070001@v.loewis.de> >> Because the users of the service have actually requested that it works >> this way. > > I've not seen this. Citations, please? Most recently on this list: Rene Dudfield. > Do you have > some secret feedback channel I could get some access to? Actually, being one of the two maintainers of the software, I do get comments in private email also, or in person (but of course you can't get access to *that*) > 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? Of course I see a difference - I was just refuting your claim that "PyPI is a catalog. [...] Anything more dilutes the purpose and turns it into something political." As for forcing features onto authors - there have been vocal requests to reject packages from PyPI that don't also host their files on it. Regards, Martin From jim at zope.com Tue Apr 5 23:33:09 2011 From: jim at zope.com (Jim Fulton) Date: Tue, 5 Apr 2011 17:33:09 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B834B.4070109@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 5:02 PM, "Martin v. L?wis" wrote: >> I don't think ratings are helpful to Django users. Do you know of any >> packages for which you think the ratings we have now actually help >> users? > > I'd have to search through the package list - but yes, I do think I'd > find examples. OK, a friendly challenge. Find one. :) (Or maybe someone else who thinks ratings are useful can find one.) Jim -- Jim Fulton http://www.linkedin.com/in/jimfulton From jacob at jacobian.org Tue Apr 5 23:38:46 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 16:38:46 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B8965.1070001@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 4:28 PM, "Martin v. L?wis" wrote: > Most recently on this list: Rene Dudfield. OK, so we're up to: * Two votes +1 (I'm assuming I can count you as a +1) * One vote in reply to me on twitter +0. * Six votes on this list -1 * Nine votes in reply to me on Twitter -1 (actually just got a tenth, but I'll leave that out). >> Do you have >> some secret feedback channel I could get some access to? > > Actually, being one of the two maintainers of the software, > I do get comments in private email also, or in person > (but of course you can't get access to *that*) Can you quantify, then, how many people are emailing you and telling you that the feature is really useful and that they're glad you added it? Or asking you to add it in the first place? > As for forcing features onto authors - there have been vocal requests > to reject packages from PyPI that don't also host their files on it. OK, but now you're making my argument for me. I'm saying that PyPI should not force requirements on packages authors (beyond what's required technically like a name and such, and what's required legally). It sounds like you agree, at least here. So what's different about ratings? Jacob From martin at v.loewis.de Tue Apr 5 23:42:29 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:42:29 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> Message-ID: <4D9B8CC5.3040701@v.loewis.de> To put some objective data into the discussion: here is the number of ratings that got posted recently, by month: 2010 | 9 | 41 2010 | 10 | 51 2010 | 11 | 42 2010 | 12 | 51 2011 | 1 | 76 2011 | 2 | 71 2011 | 3 | 47 Regards, Martin From ubernostrum at gmail.com Tue Apr 5 23:54:30 2011 From: ubernostrum at gmail.com (James Bennett) Date: Tue, 5 Apr 2011 16:54:30 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 4:38 PM, Jacob Kaplan-Moss wrote: > On Tue, Apr 5, 2011 at 4:28 PM, "Martin v. L?wis" wrote: >> Most recently on this list: Rene Dudfield. > > OK, so we're up to: TBH I feel like the relevant question is this: Martin: is there any argument, any collection of arguments, any response from users or collection of responses from users that could be presented that would convince you to turn off ratings? If yes, what criteria would you use? What sort of arguments or responses would be needed? If no, there's not really a point to this discussion. -- "Bureaucrat Conrad, you are technically correct -- the best kind of correct." From martin at v.loewis.de Tue Apr 5 23:54:42 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 05 Apr 2011 23:54:42 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> Message-ID: <4D9B8FA2.9080200@v.loewis.de> Am 05.04.2011 23:33, schrieb Jim Fulton: > On Tue, Apr 5, 2011 at 5:02 PM, "Martin v. L?wis" wrote: >>> I don't think ratings are helpful to Django users. Do you know of any >>> packages for which you think the ratings we have now actually help >>> users? >> >> I'd have to search through the package list - but yes, I do think I'd >> find examples. > > OK, a friendly challenge. Find one. :) As a positive evaluation, consider http://pypi.python.org/pypi/basen It's a small package; based on the comments, I would trust that it works properly. As a negative evaluation, consider http://pypi.python.org/pypi/python-cjson The author is clearly not happy with the evaluation, but based on the number of comments supporting the complaints, I'd rather stay away from it. Regards, Martin From ben+python at benfinney.id.au Wed Apr 6 00:03:46 2011 From: ben+python at benfinney.id.au (Ben Finney) Date: Wed, 06 Apr 2011 08:03:46 +1000 Subject: [Catalog-sig] Please turn off ratings References: Message-ID: <87d3l0gxtp.fsf@benfinney.id.au> Jacob Kaplan-Moss writes: > I'd like to ask, again, to turn off ratings on PyPI packages. [?] > PyPI is a catalog. Naive rating/voting features aren't part of that > mission. Can we please turn this ill-considered, useless feature off? +1 -- \ ?Are you pondering what I'm pondering?? ?Umm, I think so, Don | `\ Cerebro, but, umm, why would Sophia Loren do a musical?? | _o__) ?_Pinky and The Brain_ | Ben Finney From martin at v.loewis.de Wed Apr 6 00:06:35 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 00:06:35 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> Message-ID: <4D9B926B.70300@v.loewis.de> Am 05.04.2011 23:54, schrieb James Bennett: > On Tue, Apr 5, 2011 at 4:38 PM, Jacob Kaplan-Moss wrote: >> On Tue, Apr 5, 2011 at 4:28 PM, "Martin v. L?wis" wrote: >>> Most recently on this list: Rene Dudfield. >> >> OK, so we're up to: > > TBH I feel like the relevant question is this: > > Martin: is there any argument, any collection of arguments, any > response from users or collection of responses from users that could > be presented that would convince you to turn off ratings? I could be convinced by a majority vote, in an open poll. I know Jacob is skeptical about polls, and asks that the maintainers should channel the user's voices - which I believe I'm actually doing. Turning this around: what is are the objective arguments *for* turning it off? Going by Jacob's initial list of arguments, I think they all are not good reasons to turn them off: - it's not helpful to the maintainer may be true, but that's not a reason to turn it off - maintainers should just ignore them if they are not helpful - it's not helpful to users I believe this is not true - it's not helpful the commenter why does he go through the effort, then? It must be helpful, even if just to vent frustration, or to share joy - it's not helpful to PyPI none of PyPI is helpful to PyPI. PyPI is helpful to the Python community (including this feature) Regards, Martin From jim at zope.com Wed Apr 6 00:07:42 2011 From: jim at zope.com (Jim Fulton) Date: Tue, 5 Apr 2011 18:07:42 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B8FA2.9080200@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8FA2.9080200@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 5:54 PM, "Martin v. L?wis" wrote: > Am 05.04.2011 23:33, schrieb Jim Fulton: >> On Tue, Apr 5, 2011 at 5:02 PM, "Martin v. L?wis" wrote: >>>> I don't think ratings are helpful to Django users. Do you know of any >>>> packages for which you think the ratings we have now actually help >>>> users? >>> >>> I'd have to search through the package list - but yes, I do think I'd >>> find examples. >> >> OK, a friendly challenge. Find one. :) > > As a positive evaluation, consider > > http://pypi.python.org/pypi/basen > > It's a small package; based on the comments, I would trust that it > works properly. > > As a negative evaluation, consider > > http://pypi.python.org/pypi/python-cjson > > The author is clearly not happy with the evaluation, but based on the > number of comments supporting the complaints, I'd rather stay away from > it. Those look like good examples. They clearly, to me anyway, show that ratings, especially with comments can provide value to people looking for packages. Obviously, it's fair to ask whether the good outweighs the bad. Jim -- Jim Fulton http://www.linkedin.com/in/jimfulton From martin at v.loewis.de Wed Apr 6 00:24:51 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 00:24:51 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8FA2.9080200@v.loewis.de> Message-ID: <4D9B96B3.2090403@v.loewis.de> > Those look like good examples. They clearly, to me anyway, show that > ratings, especially with comments can provide value to people looking > for packages. > > Obviously, it's fair to ask whether the good outweighs the bad. Indeed. I believe that package authors and package users have diametrical desires in this matter, and that catalog-sig leans towards the author's side of the things. There have been objective complaints about the comment feature - namely, that it puts a burden on the authors to consider another support channel. This concern is now resolved by authors being able to opt out of comments. I wonder what other objective complaints are: "unfair press" can't be it - users are well able to evaluate the comments also, and judge whether 2 ratings on package are relevant or not. In any case, I would be willing to move this off-site (preferably if somebody is willing to maintain it as a separate infrastructure), as long as I am allowed to direct users to this offsite service. Regards, Martin From jjl at pobox.com Tue Apr 5 23:51:57 2011 From: jjl at pobox.com (John J Lee) Date: Tue, 5 Apr 2011 22:51:57 +0100 (BST) Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B8965.1070001@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> Message-ID: On Tue, 5 Apr 2011, "Martin v. L?wis" wrote: [...] > Actually, being one of the two maintainers of the software, > I do get comments in private email also, or in person > (but of course you can't get access to *that*) [...] Do the users you hear from cite other software rating systems that work well? What are those systems? Was the PyPI rating system modelled in any way on other rating systems? Apologies if you've answered those questions before. John From martin at v.loewis.de Wed Apr 6 00:44:53 2011 From: martin at v.loewis.de (=?ISO-8859-15?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 00:44:53 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> Message-ID: <4D9B9B65.1070507@v.loewis.de> Am 05.04.2011 23:51, schrieb John J Lee: > On Tue, 5 Apr 2011, "Martin v. L?wis" wrote: > [...] >> Actually, being one of the two maintainers of the software, >> I do get comments in private email also, or in person >> (but of course you can't get access to *that*) > [...] > > Do the users you hear from cite other software rating systems that work > well? It's been a while, but I think people referred to other rating systems in general, not to software rating. > Was the PyPI rating system modelled in any way on other rating systems? Yes, amazon.com (although it doesn't implement many of the amazon review system features; I think amazon has also grown new features since I implemented PyPI commenting). Regards, Martin From pje at telecommunity.com Wed Apr 6 01:03:08 2011 From: pje at telecommunity.com (P.J. Eby) Date: Tue, 05 Apr 2011 19:03:08 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B7EB7.3070207@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> Message-ID: <20110405230319.08A133A4063@sparrow.telecommunity.com> At 10:42 PM 4/5/2011 +0200, Martin v. L?wis wrote: > >> In any case, the community had agreed to to allow rating but to > >> make commenting optional (at the package author's choice). I don't > >> see what has changed since. > > > > Who is this "community" you speak of, and when and where did "they" > > agree? > >We performed a user poll, which came with the status quo. IIRC, that poll had a poor design and a biased interpretation. Properly interpreted, the vast majority of votes were for *other* solutions than the status quo. (In other words, the apparent superiority of this choice was an illusion created by the "no" votes being split between other alternatives.) From martin at v.loewis.de Wed Apr 6 01:08:39 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 01:08:39 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <20110405230319.08A133A4063@sparrow.telecommunity.com> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> Message-ID: <4D9BA0F7.1060500@v.loewis.de> Am 06.04.2011 01:03, schrieb P.J. Eby: > At 10:42 PM 4/5/2011 +0200, Martin v. L?wis wrote: >> >> In any case, the community had agreed to to allow rating but to >> >> make commenting optional (at the package author's choice). I don't >> >> see what has changed since. >> > >> > Who is this "community" you speak of, and when and where did "they" >> > agree? >> >> We performed a user poll, which came with the status quo. > > IIRC, that poll had a poor design and a biased interpretation. That may well be. I'm willing to run another one, and use any ballot that this list comes up with (as long as the outcome would be accepted by all). Regards, Martin From lac at openend.se Wed Apr 6 01:25:06 2011 From: lac at openend.se (Laura Creighton) Date: Wed, 06 Apr 2011 01:25:06 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: Message from =?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?= of "Wed, 06 Apr 2011 00:06:35 +0200." <4D9B926B.70300@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> <4D9B926B.70300@v.loewis.de> Message-ID: <201104052325.p35NP6Ct004630@theraft.openend.se> In a message of Wed, 06 Apr 2011 00:06:35 +0200, "Martin v. L?wis" writes: >- it's not helpful to users > I believe this is not true This is the one I am interested in changing your opinion about. I think that 'I want a 5 star rating system so I can tell what package to use without going to the effort of evaluating the system' fits right in there with 'I want microbenchmarks to tell me how fast my program will run' and 'I want taking a glass of Orange Juice every morning to prevent me from getting colds'. They all are a form of 'I want a magic bullet' where there *is* no magic bullet. If you want to find out information about a package, you need to either do the research, or read detailed comments from people who have done the research, and whose opinions you trust. Crowdsourcing does not cut it here. Anybody can rate something, and since as you say "why does he go through the effort, then? It must be helpful, even if just to vent frustration, or to share joy" you are tacitly admitting that what the rating system best measures is people's emotional reaction to your software. But this is a spectacularly poor thing to use as a measure of whether you should use somebody's software or not. It totally misses the difficult questions. And indeed, when you looked for a poorly rated package, and game up with python-cjson, it wasn't just the rating that you looked at, but the comments as well which helped you decide that this was a package you would want to stay away from. Without that information, you too couldn't tell what the rating meant. I was talking with the people at Yelp, the ratings company in San Francisco this March. They are crowdsourcing experts. And they say that there is a particular problem with people giving ratings when they don't also come with a review. People really do go around blasting other people's businesses because they think it will make their competing business prosper. People really do get drunk and decide that trashing the reputations of famous whatevers would be a really cool thing to do. People rate things poorly, because they were feeling bad at the time, for reasons that have nothing to do with the product or service. Indeed yelp has a whole department of people whose complete job is going back and correcting the ratings of people and places, by weeding out those complaints which really were only about people expressing their emotions. People using a rating system as an emotional dumping ground is a real problem for their business. And, among other things that they told me, the smaller your sample set, the worse this problem is for the rating organisation. Now, of course, Yelp has to work very hard at this, because their business model depends on people trusting the ratings that they get from Yelp when they want to purchase a good or service. So quality control for the rating system, to make sure that it is as fair as they can make it, is something they really can spend time and money at. We don't have those kinds of resources, and we don't have the kinds of numbers of reviews where a policy of 'cross your fingers and hope that the large number of fair, unbiased votes will drown out the unfair votes' can be expected to work, even in the short term. This is because the whole voting system is a self-selcted set, and the voters are disproportionately going to be the people who want to express frustration and who just like to cause trouble by giving out negative votes for the fun of annoying people, or promoting their competing package, or whatever. Thus we have a situation here where we are in more desparate need of a way to check that the ratings are fair, no resources to do so, and, given than people give votes without needing to state a reason or an explanation for their number, no way to check that the rating is fair. All we have done is undermined the credibility of http://pypi.python.org/pypi - yes there is a rating system, but there is no quality control of the raters, so its all pretty meaningless. So no magic bullet for the people who hoped that a rating system would let them pick software. But they will undoubtably try to use the ratings for exactly that, regardless, demonstrating that hope still springs eternal in the human soul. But maybe we have a responsibility to these people, to not give them what they asked for, because it really won't do what they want? It's undoubtablly fun for the people who want an emotional dumping ground, but I for one am not interested in using the catalog for these people's emotional needs. Writing software is hard enough without adding the requirement -- oh, yes, and if you ever release this thing, you have to put up with having people vent vent their frustrations with you, publicly, permanently, where there is nothing you can do about it. If you complain you get told to ignore the ratings, which it tantamount to admitting that they are meaningless. I don't want this sort of vulnerability, and I am pretty thick-skinned. It must be excruciating hell for those developers who are more senstive to public opinion than I am. So I think that the rating system is a serious disservice to the people it was supposed to help, the users who asked for it, as well as being a source of considerable angst for the software developers. Laura From pje at telecommunity.com Wed Apr 6 01:40:32 2011 From: pje at telecommunity.com (P.J. Eby) Date: Tue, 05 Apr 2011 19:40:32 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B9B65.1070507@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> <4D9B9B65.1070507@v.loewis.de> Message-ID: <20110405234053.EA6DC3A4063@sparrow.telecommunity.com> At 12:44 AM 4/6/2011 +0200, Martin v. L?wis wrote: >Am 05.04.2011 23:51, schrieb John J Lee: > > On Tue, 5 Apr 2011, "Martin v. L?wis" wrote: > > [...] > >> Actually, being one of the two maintainers of the software, > >> I do get comments in private email also, or in person > >> (but of course you can't get access to *that*) > > [...] > > > > Do the users you hear from cite other software rating systems that work > > well? > >It's been a while, but I think people referred to other rating systems >in general, not to software rating. > > > Was the PyPI rating system modelled in any way on other rating systems? > >Yes, amazon.com (although it doesn't implement many of the amazon review >system features; I think amazon has also grown new features since I >implemented PyPI commenting). One important feature missing is that the Amazon review system has *always* had a back-channel for removing unhelpful reviews; in the beginning, it was simply the ability for authors and publishers to ask Amazon to remove them, and later, the "did you find this helpful?" buttons were added. Another important way that Amazon reviews differ is that they're called *reviews*, not comments, and reviewers are rewarded in various ways for making reviews -- they are actually scored on the number and usefulness of their reviews, which encourages high-quality reviews. Good Amazon reviewers typically give both the pros and cons of a product, and describe their use case so others know whether the review is applicable to them or not. This is socially reinforced by example and by the rewards machinery. In contrast, PyPI's comment system much more closely resembles YouTube in its physical and social structure for comments... and likewise, the comments it gets more closely resemble YouTube comments than they do Amazon reviews. From martin at v.loewis.de Wed Apr 6 01:45:13 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 01:45:13 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <201104052325.p35NP6Ct004630@theraft.openend.se> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> <4D9B926B.70300@v.loewis.de> <201104052325.p35NP6Ct004630@theraft.openend.se> Message-ID: <4D9BA989.8080406@v.loewis.de> > I was talking with the people at Yelp, the ratings company in San Francisco > this March. They are crowdsourcing experts. And they say that there is > a particular problem with people giving ratings when they don't also come > with a review. I certainly agree with that, and that's why the PyPI rating system had the comments feature from the beginning. There was objection to that, on reasonable grounds, so it's opt-out now. I would certainly agree that it better wasn't opt-out, but alas, package authors disagree. It's their choice. > Thus we have a situation here where we are in more desparate need of > a way to check that the ratings are fair, no resources to do so, and, > given than people give votes without needing to state a reason or > an explanation for their number, no way to check that the rating is > fair. I believe that anybody remotely familiar with the internet is aware of this. *Of course* ratings may be misleading, and I'll use some software anyway despite everybody telling me that Apache is crap and Nginx is the one and only true web server that any sane person would use. > All we have done is undermined the credibility of > http://pypi.python.org/pypi - yes there is a rating system, but there > is no quality control of the raters, so its all pretty meaningless. User's aren't so dumb to not know that. They will read the reviews if present, and the rebuttals of the reviews if present. If there are overwhelmingly negative reviews, and no comments even though the reviewers could have commented, they assume gaming. If there are overwhelmingly negative reviews and commenting disabled, they assume the author has something to hide. If there is a single negative review, they download Django, anyway. > So I think that the rating system is a serious disservice to the people > it was supposed to help, the users who asked for it, as well as being > a source of considerable angst for the software developers. I know that you believe that, and I still disagree. Regards, Martin From jacob at jacobian.org Wed Apr 6 01:48:43 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 18:48:43 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> Message-ID: [Oops, forgot to reply all - sorry for the duplicate, Martin!] On Tue, Apr 5, 2011 at 4:42 PM, "Martin v. L?wis" wrote: > To put some objective data into the discussion: here is the number of > ratings that got posted recently, by month: > > 2010 | 9 | 41 > 2010 | 10 | 51 > 2010 | 11 | 42 > 2010 | 12 | 51 > 2011 | 1 | 76 > 2011 | 2 | 71 > 2011 | 3 | 47 Thank you. Here are some more numbers: I took a look [1] at how many ratings packages have received. Here's the distribution (# of ratings: # of packages): 0: 13452 1: 427 2: 91 3: 22 4: 12 5: 10 6: 3 7: 8 8: 4 9: 2 10: 1 11: 2 13: 1 21: 1 So 96% of packages have no ratings. Only 1% have more than one rating. I ask again, how is this useful? Django, with 7 ratings, is in the 99.9th percentile among all packages, so if anyone could find this useful I suppose it might be me. But 7 ratings isn't in any way a significant sample: PyPI shows 11,000 downloads of Django 1.3 alone (which is just a few weeks old); there's no *way* this provides *any* useful data whatsoever. Jacob [1] Using this script:: from collections import defaultdict import xmlrpclib pypi = xmlrpclib.ServerProxy('http://pypi.python.org/pypi') all_packages = pypi.list_packages() all_counts = defaultdict(int) for p in all_packages: ratings = pypi.ratings(p, "", 0) print p, len(ratings[0]) all_counts[len(ratings[0])] += 1 open('counts.txt', 'w').write(str(all_counts)) print all_counts From martin at v.loewis.de Wed Apr 6 01:50:26 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 01:50:26 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <20110405234053.EA6DC3A4063@sparrow.telecommunity.com> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> <4D9B9B65.1070507@v.loewis.de> <20110405234053.EA6DC3A4063@sparrow.telecommunity.com> Message-ID: <4D9BAAC2.4050104@v.loewis.de> > One important feature missing is that the Amazon review system has > *always* had a back-channel for removing unhelpful reviews; in the > beginning, it was simply the ability for authors and publishers to ask > Amazon to remove them, and later, the "did you find this helpful?" > buttons were added. I'd be willing to add that - although I doubt it would make Jacob happy. In any case, it's certainly possible to ask for the removal of outright spam. Regular "unhelpful" reviews shouldn't be removed, IMO, but rather marked as such - which is possible through the option of rebuttal. > In contrast, PyPI's comment system much more closely resembles YouTube > in its physical and social structure for comments... and likewise, the > comments it gets more closely resemble YouTube comments than they do > Amazon reviews. These are all important things to consider, and I'd be willing to change it, assuming catalog-sig would then accept the feature, instead of challenging it every two years. Regards, Martin From jacob at jacobian.org Wed Apr 6 01:51:29 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 18:51:29 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9BA0F7.1060500@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 6:08 PM, "Martin v. L?wis" wrote: > That may well be. I'm willing to run another one, and use any ballot > that this list comes up with (as long as the outcome would be accepted > by all). I asked before: """ How is [Pypi] any different from any other open source project? The majority of Python users don't subscribe to python-dev, yet the core Python team makes decisions on their behalf every day. As maintainers we rely on our ability to "channel" the needs of our community and act according. """ I think you're hearing a clear message from users here, from the people I quoted from Twitter -- I can copy over a dozen more replies if you like -- and from other sources that this feature is ill-considered and shouldn't be part of PyPI. Why are we pretending this is a democracy? You're clearly the BDFL as far as PyPI's concerned, so it's time make a call. Tell me to shut up, and I'll shut up. Or make the change. Just make a call and be done with it. Jacob From martin at v.loewis.de Wed Apr 6 02:00:54 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 02:00:54 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> Message-ID: <4D9BAD36.3010002@v.loewis.de> > """ > How is [Pypi] any different from any other open source project? The > majority of Python users don't subscribe to python-dev, yet the core > Python team makes decisions on their behalf every day. As maintainers > we rely on our ability to "channel" the needs of our community and act > according. > """ > > I think you're hearing a clear message from users here Sorry, no, I don't. I hear a loud message, but not a clear one (i.e. I think it comes from a vocal minority) > You're clearly the BDFL as far as PyPI's concerned, so it's time make > a call. Tell me to shut up, and I'll shut up. Or make the change. Just > make a call and be done with it. If that's what it takes: I see no reason to disable the feature at this point - but I'm willing to adjust it if we can arrive at some consensus what change to make. Regards, Martin From jacob at jacobian.org Wed Apr 6 02:10:10 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 19:10:10 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9BAD36.3010002@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 7:00 PM, "Martin v. L?wis" wrote: > Sorry, no, I don't. I hear a loud message, but not a clear one > (i.e. I think it comes from a vocal minority) I've given you loads of data including usage statistics, anecdotal data, user polling on Twitter (and I could add a link to Reddit now, but why bother). You've heard from dozens of people who're very invested in our community that this feature makes us look bad. You've countered it with a single user who likes the feature and claims of support in private email. Neither of us can prove we're speaking for a larger community. You have root so you get to rule by fiat, and I'm going to accept it and shut up, but don't you dare accuse me of representing a vocal minority when you're representing an even smaller one. This entire process leaves a terrible taste in my mouth. I hope you'll seriously consider what sort of message this governance model sends to the community. It leaves me absolutely unsure that PyPI represents my needs as a Python user. Jacob From martin at v.loewis.de Wed Apr 6 02:33:15 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 02:33:15 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: <4D9BB4CB.2030506@v.loewis.de> > This entire process leaves a terrible taste in my mouth. I hope you'll > seriously consider what sort of message this governance model sends to > the community. It leaves me absolutely unsure that PyPI represents my > needs as a Python user. What IMO made this discussion unsuccessful is that you seem to be unwilling to compromise. I can accept that the specific implementation is ill-designed, and I'm willing to change it, and have offered a number of alternatives - all of which you have ignored. Apparently, you cannot accept anything else but the total removal of the feature, which I honestly think doesn't reflect the needs of the majority of the Python users. Regards, Martin From ubernostrum at gmail.com Wed Apr 6 03:13:33 2011 From: ubernostrum at gmail.com (James Bennett) Date: Tue, 5 Apr 2011 20:13:33 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9BB4CB.2030506@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <4D9BB4CB.2030506@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 7:33 PM, "Martin v. L?wis" wrote: > What IMO made this discussion unsuccessful is that you seem to be > unwilling to compromise. I can accept that the specific implementation > is ill-designed, and I'm willing to change it, and have offered a number > of alternatives - all of which you have ignored. What made this unsuccessful, and what made the last round unsuccessful, and what will make the next round unsuccessful, and the one after that, etc., is a couple of simple facts: 1. The only evidence presented for benefits of ratings is a couple of anecdotes and some vague hand-waving about people who've emailed you, but who won't be identified or quoted. 2. The people who are, right now, presenting you with reasonable criticism and asking you to disable this feature are the very people who make PyPI worthwhile: the package maintainers. Our concerns are just as valid as anyone's, and without us PyPI is a blank web page. But we get ignored in favor of the anonymous masses who -- you claim -- support, want and dearly love this feature despite basically never actually using it. This is not reasonable and this is not acceptable, and is not the first time we've reached this impasse. What is required here is for you to accept that package maintainers can actually raise valid concerns, and that being responsive to those concerns and accommodating them is vitally necessary for PyPI's continued existence. If you're unable to do that, please let us know so that steps can be taken to deal with the problem. -- "Bureaucrat Conrad, you are technically correct -- the best kind of correct." From lists at zopyx.com Wed Apr 6 03:29:36 2011 From: lists at zopyx.com (Andreas Jung) Date: Wed, 06 Apr 2011 03:29:36 +0200 Subject: [Catalog-sig] Please turn off ratings Message-ID: <1302053376.670732@smtp.zopyx.net> > Hi folks -- > > Just got another one of these: > > > From: PyPI operators > > Subject: New rating on Django > > > > [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] > > xen has rated your package as 0/5. Comment (optional): > > I'd like to ask, again, to turn off ratings on PyPI packages. > > How is this feature helpful to anyone at all? > > * It's not helpful to me as a maintainer: I have no idea *why* "xen" > gave this rating. > * It's not helpful to users: they have no idea what "0" means. > * It's not helpful to "xen": his feedback can't be acted upon, so > we'll never be able to aprove. > * It's not helpful to PyPI: what value does "0/5" provide to a catalog? > * It's not helpful to the Python language: how does "0/5" in any way > help Python users choose tools? > > PyPI is a catalog. Naive rating/voting features aren't part of that > mission. Can we please turn this ill-considered, useless feature off? > > -1 I want to have an option to leave a comment for the public on case of bad or broken package especially if the author is not reachable or responding e.g. due to invalid or improper email addresses. leaving a comment and a vote is valid measure to deal with trash packages on pypi. yes, this discussion again .... pypi is a catalog but it should not be a package dumpster (not related to django in any way). Regards, Andreas Jung -- Sorry for being brief - sent by iPhone From richard at python.org Wed Apr 6 03:29:46 2011 From: richard at python.org (Richard Jones) Date: Wed, 6 Apr 2011 11:29:46 +1000 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> Message-ID: On Wed, Apr 6, 2011 at 9:48 AM, Jacob Kaplan-Moss wrote: > On Tue, Apr 5, 2011 at 4:42 PM, "Martin v. L?wis" wrote: >> To put some objective data into the discussion: here is the number of >> ratings that got posted recently, by month: >> >> ?2010 | ?9 | ? ?41 >> ?2010 | 10 | ? ?51 >> ?2010 | 11 | ? ?42 >> ?2010 | 12 | ? ?51 >> ?2011 | ?1 | ? ?76 >> ?2011 | ?2 | ? ?71 >> ?2011 | ?3 | ? ?47 > > Thank you. > > Here are some more numbers: I took a look [1] at how many ratings > packages have received. Here's the distribution (# of ratings: # of > packages): > > 0: 13452 > 1: 427 > 2: 91 > 3: 22 > 4: 12 > 5: 10 > 6: 3 > 7: 8 > 8: 4 > 9: 2 > 10: 1 > 11: 2 > 13: 1 > 21: 1 > > So 96% of packages have no ratings. Only 1% have more than one rating. > > I ask again, how is this useful? > > Django, with 7 ratings, is in the 99.9th percentile among all > packages, so if anyone could find this useful I suppose it might be > me. But 7 ratings isn't in any way a significant sample: PyPI shows > 11,000 downloads of Django 1.3 alone (which is just a few weeks old); > there's no *way* this provides *any* useful data whatsoever. It seems to me that, given we're quite a long way into the potential user adoption of this feature, it's simply not being used. I would support removing it, as it seems that the only purpose it serves is to antagonise. And we could all be doing far more interesting things. Richard From jacob at jacobian.org Wed Apr 6 04:10:54 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 5 Apr 2011 21:10:54 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9BB4CB.2030506@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <4D9BB4CB.2030506@v.loewis.de> Message-ID: On Tue, Apr 5, 2011 at 7:33 PM, "Martin v. L?wis" wrote: > What IMO made this discussion unsuccessful is that you seem to be > unwilling to compromise. That's a fair enough criticism and I'll own it. I have been -- and still am -- unwilling to compromise. I accept that such an attitude is difficult to listen to, and I apologize for the part my unwillingness to compromise played in failing to find a resolution. For your part, though, I hope that you'll accept that your stubbornness has more or less matched mine here. You have a preconceived notion of the usefulness of this feature and have refused to address any information that contradicts that notion. Similarly, you seem to believe that you speak on behalf of PyPI's users but can only cite private emails as evidence and ignore any public evidence to the contrary. If you've indeed been listening to my concerns you've done a poor job indicating that you hear me. > I can accept that the specific implementation > is ill-designed, and I'm willing to change it, and have offered a number > of alternatives - all of which you have ignored. What!? That's a completely unfair characterization of the conversation. You have no way of knowing that I've "ignored" your alternatives. In fact, I've read every single word of it, and indeed responded to at least one of your proposals in detail. You wrote: > If you can suggest a procedure that fairly > involves the end users also (which *frequently* had requested from > me that I provide this very feature), and they now say they don't > like this, I'd be willing to reconsider. To which I replied: > So something simple would be to let package authors provide a "feedback" link. I took your lack of response to mean that you think my suggestion is a bad idea, and fair enough. But I'm not going to characterize your lack of response as ignoring me. It's incredibly insulting to allege that I'm ignoring your suggestions. > Apparently, you cannot accept anything else but the total removal of > the feature, which I honestly think doesn't reflect the needs of the > majority of the Python users. We're back to this, again: you have this firm belief that you speak for the majority of Python users. Well, I believe the same: I'm speaking on the behalf of the majority of Python users, too. Of course, neither of us can prove that assertion. Jacob From ben+python at benfinney.id.au Wed Apr 6 06:31:55 2011 From: ben+python at benfinney.id.au (Ben Finney) Date: Wed, 06 Apr 2011 14:31:55 +1000 Subject: [Catalog-sig] Please turn off ratings References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> <4D9B9B65.1070507@v.loewis.de> <20110405234053.EA6DC3A4063@sparrow.telecommunity.com> Message-ID: <874o6cgfus.fsf@benfinney.id.au> "P.J. Eby" writes: > At 12:44 AM 4/6/2011 +0200, Martin v. L?wis wrote: > >Am 05.04.2011 23:51, schrieb John J Lee: > > > Was the PyPI rating system modelled in any way on other rating > > > systems? > > > >Yes, amazon.com (although it doesn't implement many of the amazon > >review system features; I think amazon has also grown new features > >since I implemented PyPI commenting). > > One important feature missing is that the Amazon review system has > *always* had a back-channel for removing unhelpful reviews; in the > beginning, it was simply the ability for authors and publishers to ask > Amazon to remove them, and later, the "did you find this helpful?" > buttons were added. > > Another important way that Amazon reviews differ is that they're > called *reviews*, not comments, and reviewers are rewarded in various > ways for making reviews -- they are actually scored on the number and > usefulness of their reviews, which encourages high-quality reviews. Yet another important difference: Amazon reviews are associated with a product that *won't change* for the lifetime of that review. PyPI comments and ratings, on the other hand, stick around while the package continues to accrue new versions ? so there's little way of knowing whether the comments and ratings are still relevant to the latest version of the package. Now, one possible response to that is to make the comments and rating system more complicated to account for the fact that software packages are much more dynamic than Amazon products. I hope that's not the response that is chosen. -- \ ?? whoever claims any right that he is unwilling to accord to | `\ his fellow-men is dishonest and infamous.? ?Robert G. | _o__) Ingersoll, _The Liberty of Man, Woman and Child_, 1877 | Ben Finney From monitor at jacobian.org Wed Apr 6 07:06:55 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 00:06:55 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection failed Message-ID: <1302066420.1@jacobian.org> Connection failed Service pypi.python.org Date: Wed, 06 Apr 2011 00:06:55 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[pypi.python.org:80] via TCP Your faithful employee, monit From tjreedy at udel.edu Wed Apr 6 08:13:26 2011 From: tjreedy at udel.edu (Terry Reedy) Date: Wed, 06 Apr 2011 02:13:26 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: On 4/5/2011 8:10 PM, Jacob Kaplan-Moss wrote: > On Tue, Apr 5, 2011 at 7:00 PM, "Martin v. L?wis" wrote: >> Sorry, no, I don't. I hear a loud message, but not a clear one >> (i.e. I think it comes from a vocal minority) > > I've given you loads of data including usage statistics, anecdotal > data, user polling on Twitter (and I could add a link to Reddit now, But it seems to me likely biased, so I am not convinced yet. > but why bother). You've heard from dozens of people who're very > invested in our community that this feature makes us look bad. > > You've countered it with a single user who likes the feature and > claims of support in private email. No, he countered with a community poll that has, as I remember, 100-200 responders. After a contentious discussion. > Neither of us can prove we're speaking for a larger community. You > have root so you get to rule by fiat, and I'm going to accept it and > shut up, but don't you dare accuse me of representing a vocal minority > when you're representing an even smaller one. > > This entire process leaves a terrible taste in my mouth. I agree, but for a quite different reason. Martin sometimes gets opposite demands. That is not easy to deal with. On this issue, after much discussion Martin took a poll and pretty fairly, in my opinion, followed the result. You did not like that. You get an email that to you is spam. You do not like this. I would agree that you should be able to opt out of receiving them, if you cannot do so now. An avalanche follows. It mostly consists of people who did not ratings before saying they do not like ratings now, and perhaps even less. Surprise. And there is some rehashing of the same old arguments. A small number of posts give new data and arguments. I even find them a bit persuasive. I might have found them even more persuasive if not nearly lost in the rest, which had already left a bad taste in *my* mouth. A couple of argument strike me a strange. On is the Catch-22 argument that ratings without comments are bad, when the person making that argument is the one who turned comments off. The other is the argument that Martin it is acting like a dictator because he rejects the demand that he act like a dictator by ignoring the poll. That's how it seem from here, anyhow. -- Terry Jan Reedy From chris at simplistix.co.uk Wed Apr 6 09:06:05 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 08:06:05 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: <4D9C10DD.6040709@simplistix.co.uk> On 05/04/2011 19:28, Jacob Kaplan-Moss wrote: > Hi folks -- > > Just got another one of these: > >> From: PyPI operators >> Subject: New rating on Django >> >> [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] >> xen has rated your package as 0/5. Comment (optional): > > I'd like to ask, again, to turn off ratings on PyPI packages. > > How is this feature helpful to anyone at all? > > * It's not helpful to me as a maintainer: I have no idea *why* "xen" > gave this rating. > * It's not helpful to users: they have no idea what "0" means. > * It's not helpful to "xen": his feedback can't be acted upon, so > we'll never be able to aprove. > * It's not helpful to PyPI: what value does "0/5" provide to a catalog? > * It's not helpful to the Python language: how does "0/5" in any way > help Python users choose tools? > > PyPI is a catalog. Naive rating/voting features aren't part of that > mission. Can we please turn this ill-considered, useless feature off? +1 Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From chris at simplistix.co.uk Wed Apr 6 09:08:50 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 08:08:50 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9BAAC2.4050104@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <4D9B813A.7010103@v.loewis.de> <4D9B84A9.2090904@v.loewis.de> <4D9B8965.1070001@v.loewis.de> <4D9B9B65.1070507@v.loewis.de> <20110405234053.EA6DC3A4063@sparrow.telecommunity.com> <4D9BAAC2.4050104@v.loewis.de> Message-ID: <4D9C1182.60008@simplistix.co.uk> On 06/04/2011 00:50, "Martin v. L?wis" wrote: > These are all important things to consider, and I'd be willing to change > it, assuming catalog-sig would then accept the feature, instead of > challenging it every two years. Maybe the fact that this is always coming up is a sign? cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From monitor at jacobian.org Wed Apr 6 09:10:26 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 02:10:26 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection succeeded Message-ID: <1302073829.1@jacobian.org> Connection succeeded Service pypi.python.org Date: Wed, 06 Apr 2011 02:10:26 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[pypi.python.org:80] via TCP Your faithful employee, monit From chris at simplistix.co.uk Wed Apr 6 09:12:28 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 08:12:28 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> Message-ID: <4D9C125C.9060508@simplistix.co.uk> On 06/04/2011 02:29, Richard Jones wrote: > It seems to me that, given we're quite a long way into the potential > user adoption of this feature, it's simply not being used. I would > support removing it, as it seems that the only purpose it serves is to > antagonise. And we could all be doing far more interesting things. +sys.maxint. I'm confused as to why with only one person advocating this feature, it hasn't simply been dropped. And Martin, before you cite the masses you purport to represent, please take into account the actual stats that have been presented to you. cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From renesd at gmail.com Wed Apr 6 09:12:32 2011 From: renesd at gmail.com (=?ISO-8859-1?Q?Ren=E9_Dudfield?=) Date: Wed, 6 Apr 2011 08:12:32 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9C10DD.6040709@simplistix.co.uk> References: <4D9C10DD.6040709@simplistix.co.uk> Message-ID: Can we please continue this conversation over the next week? It's been very useful. I'd like to see if we can get up to 1000 emails on this topic concerning if ratings out of 5 are good or not. -------------- next part -------------- An HTML attachment was scrubbed... URL: From chris at simplistix.co.uk Wed Apr 6 09:16:08 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 08:16:08 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9B8730.5010106@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B8730.5010106@v.loewis.de> Message-ID: <4D9C1338.6@simplistix.co.uk> On 05/04/2011 22:18, "Martin v. L?wis" wrote: > In abstract, the rationale is that it allows users to select a package > without going through the possibly-time-consuming process of evaluating. Encouraging any user, particularly new users, to select a package based on a 5-point abstract rating system and a couple of user comments is doing them a dis-service. Selecting open source packages to use is one of the more difficult tasks a developer who chooses to use them can face. Encouraging them *not* to do their own research (find the issue tracker, see what issues are open, see how active the mailing list is, see when releases were last made, see if the maintainer is responsive, etc) seems like an extremely bad idea. cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From chris at simplistix.co.uk Wed Apr 6 09:18:36 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 08:18:36 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <20110405212727.EB71F3A4063@sparrow.telecommunity.com> References: <20110405212727.EB71F3A4063@sparrow.telecommunity.com> Message-ID: <4D9C13CC.9030108@simplistix.co.uk> On 05/04/2011 22:27, P.J. Eby wrote: > At 01:28 PM 4/5/2011 -0500, Jacob Kaplan-Moss wrote: >> PyPI is a catalog. Naive rating/voting features aren't part of that >> mission. Can we please turn this ill-considered, useless feature off? > > +1 We have a switch to disable commenting. Please can we have the same for ratings? That way, those of us who pride ourselves on providing documented packages with obvious links to mailing lists, issue trackers, and source repositories can try and help our package users by stopping them being subjected to this annoying distaction. Package maintainers who don't care (nested list printer anyone?) can leave it on and potential package users can judge them as a result. cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From chris at simplistix.co.uk Wed Apr 6 09:19:41 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 08:19:41 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9C10DD.6040709@simplistix.co.uk> Message-ID: <4D9C140D.1040005@simplistix.co.uk> On 06/04/2011 08:12, Ren? Dudfield wrote: > Can we please continue this conversation over the next week? It's been > very useful. I'd like to see if we can get up to 1000 emails on this > topic concerning if ratings out of 5 are good or not. 5/5 from me. No comments, I'm too lazy... ;-) Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From ubernostrum at gmail.com Wed Apr 6 09:21:32 2011 From: ubernostrum at gmail.com (James Bennett) Date: Wed, 6 Apr 2011 02:21:32 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: On Wed, Apr 6, 2011 at 1:13 AM, Terry Reedy 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. Surprise. > 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. 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. 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 picture. 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 Ren? 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 empty. -- "Bureaucrat Conrad, you are technically correct -- the best kind of correct." From monitor at jacobian.org Wed Apr 6 10:08:17 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 03:08:17 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection failed Message-ID: <1302077302.1@jacobian.org> Connection failed Service pypi.python.org Date: Wed, 06 Apr 2011 03:08:17 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Wed Apr 6 10:43:26 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 03:43:26 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection succeeded Message-ID: <1302079410.1@jacobian.org> Connection succeeded Service pypi.python.org Date: Wed, 06 Apr 2011 03:43:26 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Wed Apr 6 10:55:33 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 03:55:33 -0500 Subject: [Catalog-sig] [monit] c.pypi.python.org - Connection failed Message-ID: <1302080136.1@jacobian.org> Connection failed Service c.pypi.python.org Date: Wed, 06 Apr 2011 03:55:33 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[c.pypi.python.org:80] via TCP Your faithful employee, monit From g.brandl at gmx.net Wed Apr 6 12:27:35 2011 From: g.brandl at gmx.net (Georg Brandl) Date: Wed, 06 Apr 2011 12:27:35 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> Message-ID: Am 06.04.2011 03:29, schrieb Richard Jones: >> So 96% of packages have no ratings. Only 1% have more than one rating. >> >> I ask again, how is this useful? >> >> Django, with 7 ratings, is in the 99.9th percentile among all >> packages, so if anyone could find this useful I suppose it might be >> me. But 7 ratings isn't in any way a significant sample: PyPI shows >> 11,000 downloads of Django 1.3 alone (which is just a few weeks old); >> there's no *way* this provides *any* useful data whatsoever. > > It seems to me that, given we're quite a long way into the potential > user adoption of this feature, it's simply not being used. I would > support removing it, as it seems that the only purpose it serves is to > antagonise. And we could all be doing far more interesting things. That seems to be a logical conclusion to me. The only thing I would keep is (if enabled for the package) the comment form, but only to give feedback to the package author(s). Georg From fuzzyman at gmail.com Wed Apr 6 13:55:45 2011 From: fuzzyman at gmail.com (Michael Foord) Date: Wed, 6 Apr 2011 12:55:45 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: On 6 April 2011 08:21, James Bennett wrote: > On Wed, Apr 6, 2011 at 1:13 AM, Terry Reedy 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. > Surprise. > > 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 raised. 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: http://mail.python.org/pipermail/python-dev/2009-November/094058.html http://mail.python.org/pipermail/python-dev/2009-November/094072.html http://mail.python.org/pipermail/python-dev/2009-November/094084.html http://mail.python.org/pipermail/python-dev/2009-November/094075.html All the best, Michael Foord > 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 > picture. > > 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 Ren? 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 > empty. > > > -- > "Bureaucrat Conrad, you are technically correct -- the best kind of > correct." > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- http://www.voidspace.org.uk/ 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... URL: From g.brandl at gmx.net Wed Apr 6 14:11:03 2011 From: g.brandl at gmx.net (Georg Brandl) Date: Wed, 06 Apr 2011 14:11:03 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: Am 06.04.2011 13:55, schrieb Michael Foord: > > > On 6 April 2011 08:21, James Bennett > wrote: > > On Wed, Apr 6, 2011 at 1:13 AM, Terry Reedy > 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. Surprise. > > 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 raised. Well, let's just find James' packages on PyPI and give them all a negative rating :) Seriously, as a lurker here I'm surprised at the lack of conduct, even if the topic is emotional. (How can a rating system be emotional, anyway? There's not even any graphical stars about whose color we could argue. I'm sure Django can cope with a negative vote.) > 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.) Great. PycebookPI anyone? :) cheers, Georg From monitor at jacobian.org Wed Apr 6 14:24:58 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 07:24:58 -0500 Subject: [Catalog-sig] [monit] c.pypi.python.org - Connection succeeded Message-ID: <1302092701.1@jacobian.org> Connection succeeded Service c.pypi.python.org Date: Wed, 06 Apr 2011 07:24:58 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[c.pypi.python.org:80] via TCP Your faithful employee, monit From tseaver at palladion.com Wed Apr 6 17:07:22 2011 From: tseaver at palladion.com (Tres Seaver) Date: Wed, 06 Apr 2011 11:07:22 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/05/2011 02:28 PM, Jacob Kaplan-Moss wrote: > Hi folks -- > > Just got another one of these: > >> From: PyPI operators >> Subject: New rating on Django >> >> [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] >> xen has rated your package as 0/5. Comment (optional): > > I'd like to ask, again, to turn off ratings on PyPI packages. > > How is this feature helpful to anyone at all? > > * It's not helpful to me as a maintainer: I have no idea *why* "xen" > gave this rating. > * It's not helpful to users: they have no idea what "0" means. > * It's not helpful to "xen": his feedback can't be acted upon, so > we'll never be able to aprove. > * It's not helpful to PyPI: what value does "0/5" provide to a catalog? > * It's not helpful to the Python language: how does "0/5" in any way > help Python users choose tools? > > PyPI is a catalog. Naive rating/voting features aren't part of that > mission. Can we please turn this ill-considered, useless feature off? +1 Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver at palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk2cgaoACgkQ+gerLs4ltQ7z8QCfXtNSjv7P/GyQTbvGu2b1o6sx 7fkAmwR4uLD1jU4UP5C8C+llHVBj7lUI =i/Zg -----END PGP SIGNATURE----- From bkjones at gmail.com Wed Apr 6 17:33:43 2011 From: bkjones at gmail.com (Brian Jones) Date: Wed, 6 Apr 2011 11:33:43 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: I think if PyPI had a ratings system that more closely matched some kind of (perhaps updated) consensus on what would be "useful", this conversation could focus more on what that looks like and who might implement it, which would be more constructive IMHO. Attempting to completely eradicate a feature by trying to collect anecdotal input doesn't seem like the way to go. Why don't we start a new thread about what a good ratings system looks like and try to get actual consensus from it so real work can get done? I, for one, am not a fan of the ratings system, but I don't agree that PyPI has no business having one. I'd be happy if someone came up with a system that was useful enough to justify its existence. Maybe there are just particular features of the existing system that make it hard or unintuitive for end users to use in a meaningful way, which then makes any resulting 'rating' useless to anyone else who happens by. Regardless, I suspect some of the vitriol toward the existing system is really about features of the system, and not about the existence of the system. Catalogs have ratings systems. I'd argue that a lot of sites that have ratings systems are, in fact, catalogs to one degree or another. Including Amazon. Jacob, what would make you *want* those emails you're getting as a package owner? What would make users *want* to leave feedback that would be useful to maintainers & other users? On Wed, Apr 6, 2011 at 11:07 AM, Tres Seaver wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On 04/05/2011 02:28 PM, Jacob Kaplan-Moss wrote: > > Hi folks -- > > > > Just got another one of these: > > > >> From: PyPI operators > >> Subject: New rating on Django > >> > >> [REPLIES TO THIS MESSAGE WILL NOT GO TO THE COMMENTER] > >> xen has rated your package as 0/5. Comment (optional): > > > > I'd like to ask, again, to turn off ratings on PyPI packages. > > > > How is this feature helpful to anyone at all? > > > > * It's not helpful to me as a maintainer: I have no idea *why* "xen" > > gave this rating. > > * It's not helpful to users: they have no idea what "0" means. > > * It's not helpful to "xen": his feedback can't be acted upon, so > > we'll never be able to aprove. > > * It's not helpful to PyPI: what value does "0/5" provide to a catalog? > > * It's not helpful to the Python language: how does "0/5" in any way > > help Python users choose tools? > > > > PyPI is a catalog. Naive rating/voting features aren't part of that > > mission. Can we please turn this ill-considered, useless feature off? > > +1 > > > Tres. > - -- > =================================================================== > Tres Seaver +1 540-429-0999 tseaver at palladion.com > Palladion Software "Excellence by Design" http://palladion.com > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.10 (GNU/Linux) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ > > iEYEARECAAYFAk2cgaoACgkQ+gerLs4ltQ7z8QCfXtNSjv7P/GyQTbvGu2b1o6sx > 7fkAmwR4uLD1jU4UP5C8C+llHVBj7lUI > =i/Zg > -----END PGP SIGNATURE----- > > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- Brian K. Jones My Blog http://www.protocolostomy.com Follow me http://twitter.com/bkjones -------------- next part -------------- An HTML attachment was scrubbed... URL: From tseaver at palladion.com Wed Apr 6 18:29:23 2011 From: tseaver at palladion.com (Tres Seaver) Date: Wed, 06 Apr 2011 12:29:23 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/06/2011 11:33 AM, Brian Jones wrote: > I think if PyPI had a ratings system that more closely matched some kind of > (perhaps updated) consensus on what would be "useful", this conversation > could focus more on what that looks like and who might implement it, which > would be more constructive IMHO. Attempting to completely eradicate a > feature by trying to collect anecdotal input doesn't seem like the way to > go. > > Why don't we start a new thread about what a good ratings system looks like > and try to get actual consensus from it so real work can get done? Why don't we take into account the information from Laura Creighton (relaying from Yelp) that getting reliable value out "crowdsourcing" requires a significant and ongoing investment of effort. Given that information, we might reconsider whether we want to deploy any half-baked version at all? I would say that systems which encourage "drive by" ratings (where ratings do not attach to the reputation of the person who creates them) are intrinsically broken. I will propose a radically simpler alternative to ratings: a simple "like this package" checkbox might go some way toward countering the bias in download counts toward packages used in repeated automated builds. I would still make the user IDs of those "likes" public, and would make each user ID a link to a page where *all* a user's "likes" are displayed. Note that I don't have any personal negative experience with PyPI's current ratings: out of the 240 PyPI projects that I (help) maintain, I couldn't find a single rating on any most-recent release. I did find a solitary 5 rating on the next-most-recent release of one package (pkginfo 0.7). Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver at palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk2clOMACgkQ+gerLs4ltQ4GqwCdEleNBxXYBNkQCCkBkK7h49vx U7YAnRNmvKgBUAl6TplEM8MjDslhu4Wz =OYAN -----END PGP SIGNATURE----- From pje at telecommunity.com Wed Apr 6 18:44:38 2011 From: pje at telecommunity.com (P.J. Eby) Date: Wed, 06 Apr 2011 12:44:38 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> Message-ID: <20110406164501.310DC3A4077@sparrow.telecommunity.com> At 02:13 AM 4/6/2011 -0400, Terry Reedy wrote: >No, he countered with a community poll that has, as I remember, >100-200 responders. After a contentious discussion. A poll which, unfortunately, had numerous choices about what *kind* of system to have, such that votes *against* Martin's proposal were widely split, and IIRC, the difference between the number of votes cast for different alternatives was pretty negligible. Looking at the numbers at one point, I concluded that if I had run a poll with those results, I would have had to conclude that there was essentially zero consensus about what direction should be taken, and either tabled the notion or gone back to the list to try to get more specific discussion. Unfortunately, the poll was run as a decision-making mechanism, rather than an information-gathering one. From fuzzyman at gmail.com Wed Apr 6 18:52:56 2011 From: fuzzyman at gmail.com (Michael Foord) Date: Wed, 6 Apr 2011 17:52:56 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <20110406164501.310DC3A4077@sparrow.telecommunity.com> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> Message-ID: On 6 April 2011 17:44, P.J. Eby wrote: > At 02:13 AM 4/6/2011 -0400, Terry Reedy wrote: > >> No, he countered with a community poll that has, as I remember, 100-200 >> responders. After a contentious discussion. >> > > A poll which, unfortunately, had numerous choices about what *kind* of > system to have, such that votes *against* Martin's proposal were widely > split, and IIRC, the difference between the number of votes cast for > different alternatives was pretty negligible. > > Looking at the numbers at one point, I concluded that if I had run a poll > with those results, I would have had to conclude that there was essentially > zero consensus about what direction should be taken, and either tabled the > notion or gone back to the list to try to get more specific discussion. > > Unfortunately, the poll was run as a decision-making mechanism, rather than > an information-gathering one. > > Well sure, but it still didn't show that *nobody* wants ratings which is what people in this thread seem to be claiming. I'm afraid that those in this thread, and with all respect to Jacob those who respond to him on twitter, *do* represent a vocal minority primarily of package authors and are not in any way representative of users of pypi. A poll may be flawed however it is done, but is a much better mechanism. Martin *has* offered to do another poll, an offer that has been ignored. All the best, Michael Foord > > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- http://www.voidspace.org.uk/ 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... URL: From jacob at jacobian.org Wed Apr 6 19:06:42 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Wed, 6 Apr 2011 12:06:42 -0500 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: On Wed, Apr 6, 2011 at 10:33 AM, Brian Jones wrote: > I, for one, am not a fan of the ratings system, but I don't agree that PyPI > has no business having one. I'd probably be less inclined to complain if the system actually did something useful, but I maintain that having a rating system is not part of the mission of a catalog, and that the existence of one, however nifty, is exclusionary, biased, and detracts from the community-resource nature of a catalog. The vast majority of programming language package catalogs don't have any sort of ratings system -- RubyGems, NPM, PEAR, Hackage, CRAN, ... in fact, the only example ( can find of a package catalog that *does* have ratings is CPAN, and those ratings are on an external site (cpanratings). I have absolutely no problem with the idea of rating packages. Heck, I built one myself once (https://github.com/jacobian/cheeserater) and I promote http://djangopackages.com/ every chance I get. I don't believe that these features are within the mission of the catalog itself, and in fact I believe they're actively harmful when implicitly endorsed by the catalogers. All that being said, it's clear that I've lost this argument, so I agree that if PyPI must have a ratings system, then it should at least be a good one. > Jacob, what would make you *want* those emails you're getting as a package > owner? What would make users *want* to leave feedback that would be useful > to maintainers & other users? Well, I might be the wrong person to ask. I'm displaying a strong bais, and I seem to be doing a great job acting like an asshole, so I suspect any suggestions coming from me are going to be dismissed out of hand at this point. That said, I'll try to engage with this question dispassionately and give a constructive answer. I'm a process nerd, so I tend to break this sort of problem down into a series of "user stories" to describe the different stakeholders and use cases a system like this should support. Here are the ones I can come up with; I think I've got most of the important ones, but perhaps I'm missing a few: 1. Amy is a Python user looking for a package to solve a specific problem. When confronted with a list of alternatives ratings could help her decide which package she should try first. When confronted with a specific package, a rating could help her decide whether the package is worth her time. For ratings to be useful to Amy, then, they'll need to specific and actionable -- that is, a rating should correspond to some indication of how "useful" the package would be. Amy needs to be able to set a threshold below which she won't consider a package. "3" doesn't help Amy: how should she decide whether a "3" is worth her time or not? "14% of users uninstalled this package within an hour" might be a more useful sort of rating for Amy. "27 people liked this package" might also be useful. A simple count of downloads is fairly useful as well. Amy needs a context against which to evaluate these ratings. Amy also needs ratings to cover a fairly broad spectrum of other users like Amy. A single rating isn't helpful, nor is just a handful. She probably needs somewhere in the neighborhood of about a dozen ratings to be fairly confidant she's getting a useful indication of utility. 2. Brian is a Python user who's downloaded a package and wants to provide feedback on it. Perhaps he's found that the tool he downloaded doesn't work and wants to complain. Perhaps he has a patch. Perhaps he's happy and wants to give the developer kudos. Like Amy, Brian needs ratings to be specific and actionable: he needs to be able to know *why* he should give a package a rating of "fresh" instead of "moldy", or what effect clicking "Like" has. Brian's been in Amy's shoes, so it's really the same problem, though for Brian, a score is more useful than a "like" since "like" won't give him any way of expressing negative feedback or patches. So Brian would appreciate other feedback mechanisms besides just a strict rating (but that's out of scope right now). 3. Carol is a package maintainer who lists her work on PyPI. Carol might want to use ratings to determine the future direction of her work, find out what users like and dislike, or even just find out if anyone's paying attention. Like Amy and Brian, Carol needs ratings to be actionable but in a slightly different manner. Positive ratings are actually fairly useless to Carol: knowing that 17 users like her code doesn't give her anything beyond an ego boost. Instead, Carol needs feedback that gives her a specific way to move forward: "fix X", "improve Y", "add Z". So numeric ratings don't particularly help Carol unless they're tied to a metric. If a score of 7.5 just represents an abstract "likeiness" rating that's useless to her, but if a score of "B-" indicates "works as advertised but has a few bugs" that's more helpful. Carol might appreciate feedback that comes on PyPI, but Carol also might have different mechanisms for users to give her feedback -- a ticket tracker, mailing list, feedback tool, etc. -- that she'd rather use. So Carol probably wants the ability to redirect or cross-post feedback into another system to avoid creating a support channel she doesn't check often. 4. Dave is a maintainer of PyPI. Dave's main goal is to help people find good code on PyPI, so Dave's interest in ratings is towards that end. Dave's thus probably most invested in seeing as many ratings as possible (along with the usual needs of making sure that the system isn't gamed or abused). Dave wants to see a rating system that provides as much information to users as possible. Thus, Dave would probably be most happy with a rating that's automatic or very easy to use. Obviousness is important to Dave: he wants users to have as little confusion around leaving a rating as possible. A "I like this" or "I use this" button is something Dave might really like: it's clear, easy to use, and has a clear precedent (Facebook). He might also be happy with systems that collect some form of "happiness" statistic or proxy automatically from users. In the end, however, Dave's main interest is just in seeing the system used and used frequently. 5. Edith is a PSF member, or board member, or director. Her main interest is similar to Dave's: she wants to be able to use PyPI as a proxy to demonstrate a vibrant, active community. She wants to point to ongoing traffic and activity on PyPI and show that Python's being used frequently and usually liked. She wants PyPI to be a place with as broad an appeal as possible -- fragmentation is her enemy. Edith has very little investment in the specifics of a ratings mechanism, though she's probably less interested in anything involving negative feedback since that sort of thing can turn into a flamewar on PyPI. But she'd probably go along with any mechanism that helps demonstrate -- and reinforce! -- the vibrancy of the Python community. ---- To me, the conclusions from the above write themselves. I think I'll avoid drawing them, though, to try to keep this as dispassionate as possible. Jacob From chris at simplistix.co.uk Wed Apr 6 19:24:13 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 18:24:13 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> Message-ID: <4D9CA1BD.30406@simplistix.co.uk> On 06/04/2011 17:52, Michael Foord wrote: > > Martin *has* offered to do another poll, an offer that has been ignored. Fine, but can it please be a fair poll this time? Do you want ratings on PyPI? Yes/No ...not a variety of options which, being cynical, are designed to split the votes of those who don't want ratings versus the people who do. cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From tseaver at palladion.com Wed Apr 6 19:24:20 2011 From: tseaver at palladion.com (Tres Seaver) Date: Wed, 06 Apr 2011 13:24:20 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/06/2011 12:52 PM, Michael Foord wrote: > On 6 April 2011 17:44, P.J. Eby wrote: > >> At 02:13 AM 4/6/2011 -0400, Terry Reedy wrote: >> >>> No, he countered with a community poll that has, as I remember, 100-200 >>> responders. After a contentious discussion. >>> >> >> A poll which, unfortunately, had numerous choices about what *kind* of >> system to have, such that votes *against* Martin's proposal were widely >> split, and IIRC, the difference between the number of votes cast for >> different alternatives was pretty negligible. >> >> Looking at the numbers at one point, I concluded that if I had run a poll >> with those results, I would have had to conclude that there was essentially >> zero consensus about what direction should be taken, and either tabled the >> notion or gone back to the list to try to get more specific discussion. >> >> Unfortunately, the poll was run as a decision-making mechanism, rather than >> an information-gathering one. >> >> > > Well sure, but it still didn't show that *nobody* wants ratings which is > what people in this thread seem to be claiming. > > I'm afraid that those in this thread, and with all respect to Jacob those > who respond to him on twitter, *do* represent a vocal minority primarily of > package authors and are not in any way representative of users of pypi. A > poll may be flawed however it is done, but is a much better mechanism. > > Martin *has* offered to do another poll, an offer that has been ignored. For those who may not be familiar with the prior poll, here are the results reported to the list[1]: - ----------------------------------- %< -------------------------------- Allow ratings and comments on all packages (status quo) 237 Allow package owners to disallow comments (ratings unmodified). 139 Allow comments, but only send them to package owners (ratings unmodified). 33 Disallow comments (ratings unmodified). 24 Disallow ratings and comments (status three months ago). 88 - ----------------------------------- %< -------------------------------- Interpretation of those results was the subject of a huge thread, which produced no clear consensus (at least to me). [1] http://permalink.gmane.org/gmane.comp.python.catalog/2169 Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver at palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk2cocQACgkQ+gerLs4ltQ5rcQCgqA1G5mw21Zvz++iMM3vc63i0 McwAoLz3aKjZ5bEg07xAcBwhHC8irMjz =caM8 -----END PGP SIGNATURE----- From pje at telecommunity.com Wed Apr 6 19:24:36 2011 From: pje at telecommunity.com (P.J. Eby) Date: Wed, 06 Apr 2011 13:24:36 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: Message-ID: <20110406172453.EE2393A4077@sparrow.telecommunity.com> At 11:33 AM 4/6/2011 -0400, Brian Jones wrote: >I think if PyPI had a ratings system that more closely matched some >kind of (perhaps updated) consensus on what would be "useful", this >conversation could focus more on what that looks like and who might >implement it, which would be more constructive IMHO. Attempting to >completely eradicate a feature by trying to collect anecdotal input >doesn't seem like the way to go. > >Why don't we start a new thread about what a good ratings system >looks like and try to get actual consensus from it so real work can get done? Better still - let's discuss the actual use cases for a rating system, before trying to design one. So far, the people who've stepped up here say they want to warn people about "bad or broken" packages, or ones that are unmaintained, or where the author is unreachable. This doesn't sound so much like a use case for a rating/comment system, as it does for a reporting mechanism for unreachable authors, or maybe something else. Side note: the ratings system's way of summarizing things doesn't make much sense either; setuptools currently rates around 2.36, with 7 people voting 0 or 1, and 7 voting 4 or 5 -- this means that the summary is equally objectionable to *both groups*... too high for the people voting low, too low for the people voting high! So, as Jacob's been saying, the number seems pretty meaningless. (This is a general problem with ratings of controversial topics, but it's especially likely given that currently, only super-exposed *and* controversial packages seem to *get* reviews at all. None of my other packages has received a single rating, AFAICT, despite many having download counts in the tens of thousands.) >I'd be happy if someone came up with a system that was useful enough >to justify its existence. Maybe there are just particular features >of the existing system that make it hard or unintuitive for end >users to use in a meaningful way, which then makes any resulting >'rating' useless to anyone else who happens by. The primary problem is that it's set up as an essentially moderator-free forum, so comment quality is low. People ignore bug reporting links and report them in comments anyway, and there's nothing the package author can do about it. One then ends up with a public bug-solving thread on the package... which then communicates to visitors that this is in fact the way to get a problem solved, thereby encouraging more of the same. In short, the social design of the feature is poor - a bad attempt at a technical solution to a fundamentally social problem. Its only saving grace is that it is *also* difficult to use (by requiring a PyPI account), so it doesn't get used very much. ;-) Note also Laura Creighton's comments regarding people wishing for ponies... the people who want comments on PyPI are essentially two groups: 1. People who want to leave negative comments 2. People who want to read other people's comments I haven't yet seen anybody say, "I want the comment feature because I want to post my in-depth evaluations and positive experiences"... which should be a good indicator of what sort of comments we should expect to see on packages. Essentially, this is because the people who actually care enough to post an informative review, will do so on their personal blogs. My guess is that, if PyPI ratings or reviews *required* a link to a blog posting, and had to be a summary or excerpt of that posting, approved by a human moderator, and with the ability for anyone (including the package author) to report that the link has become broken (and to thus get the rating/review removed), then you'd actually have a useful review/rating system. The persons who want negative reports could still write them, but they'd first have to do so on their own blog, where it's more directly associated with their online persona, and subject to other feedback mechanisms (such as comment on the blog). And it would discourage bug reports, because, well, who puts a bug report on their blog and expects to get an answer there? The big downside of course is that somebody would have to run this. It couldn't just be an unattended, hands-free system. >Regardless, I suspect some of the vitriol toward the existing system >is really about features of the system, and not about the existence >of the system. Catalogs have ratings systems. I'd argue that a lot >of sites that have ratings systems are, in fact, catalogs to one >degree or another. Including Amazon. The difference is that those systems have moderation mechanisms, and moderators... they have filters, appointed moderators, and crowd-sourced voting. And on Amazon, there are *still* people who post eBay-style feedback, "Shipped fast, would buy again" and other such crap on a product review, because they've completely missed the point of Amazon reviews. Doing a quality rating system means there needs to be somebody who: 1) actually cares about the ongoing quality of the reviews, and 2) has the ability to *do* something about it (like trimming discussions) (Oh... and on Amazon, manufacturers can post responses to reviews, but the reviewer doesn't get to respond to that, ad infinitum. It's just review+response, with an additional off-page comment system for further discussions... which usually aren't of much value as a "review" system, which is why they're off-page.) >Jacob, what would make you *want* those emails you're getting as a >package owner? What would make users *want* to leave feedback that >would be useful to maintainers & other users? I think this is mistaken -- reviews are not author feedback. As a package author, I want people to report problems to wherever I've asked them to report problems: a bug tracker or mailing list. If they want to say something nice, well, why not on their blog? From fuzzyman at gmail.com Wed Apr 6 19:25:23 2011 From: fuzzyman at gmail.com (Michael Foord) Date: Wed, 6 Apr 2011 18:25:23 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9CA1BD.30406@simplistix.co.uk> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> <4D9CA1BD.30406@simplistix.co.uk> Message-ID: On 6 April 2011 18:24, Chris Withers wrote: > On 06/04/2011 17:52, Michael Foord wrote: > >> >> Martin *has* offered to do another poll, an offer that has been ignored. >> > > Fine, but can it please be a fair poll this time? > > Do you want ratings on PyPI? Yes/No > > ...not a variety of options which, being cynical, are designed to split the > votes of those who don't want ratings versus the people who do. > > Would it be fair to split the yes vote by offering alternatives? (Not that there are any viable ones - although we could just add a Facebook like button to each package ;-) Michael > cheers, > > > Chris > > -- > Simplistix - Content Management, Batch Processing & Python Consulting > - http://www.simplistix.co.uk > -- http://www.voidspace.org.uk/ 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... URL: From chris at simplistix.co.uk Wed Apr 6 19:27:43 2011 From: chris at simplistix.co.uk (Chris Withers) Date: Wed, 06 Apr 2011 18:27:43 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> <4D9CA1BD.30406@simplistix.co.uk> Message-ID: <4D9CA28F.4030907@simplistix.co.uk> On 06/04/2011 18:25, Michael Foord wrote: > > ...not a variety of options which, being cynical, are designed to > split the votes of those who don't want ratings versus the people > who do. > > > Would it be fair to split the yes vote by offering alternatives? (Not > that there are any viable ones - although we could just add a Facebook > like button to each package ;-) It just dilutes what we're trying to establish here: do people want ratings on PyPI? Yes or no? For me, ratings belong elsewhere. As I've already stated in this thread, I believe user ratings, wherever they are, encourage lazy package picking rather than proper research by users new and experienced, which is what we *should* be encouraging. cheers, Chris -- Simplistix - Content Management, Batch Processing & Python Consulting - http://www.simplistix.co.uk From fuzzyman at gmail.com Wed Apr 6 19:29:52 2011 From: fuzzyman at gmail.com (Michael Foord) Date: Wed, 6 Apr 2011 18:29:52 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9CA28F.4030907@simplistix.co.uk> References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> <4D9CA1BD.30406@simplistix.co.uk> <4D9CA28F.4030907@simplistix.co.uk> Message-ID: On 6 April 2011 18:27, Chris Withers wrote: > On 06/04/2011 18:25, Michael Foord wrote: > >> >> ...not a variety of options which, being cynical, are designed to >> split the votes of those who don't want ratings versus the people >> who do. >> >> >> Would it be fair to split the yes vote by offering alternatives? (Not >> that there are any viable ones - although we could just add a Facebook >> like button to each package ;-) >> > > It just dilutes what we're trying to establish here: do people want ratings > on PyPI? Yes or no? > Well, it seems that this isn't the *only* thing we're trying to establish. Would another format be better is also part of the discussion. Michael > > For me, ratings belong elsewhere. As I've already stated in this thread, I > believe user ratings, wherever they are, encourage lazy package picking > rather than proper research by users new and experienced, which is what we > *should* be encouraging. > > > cheers, > > Chris > > -- > Simplistix - Content Management, Batch Processing & Python Consulting > - http://www.simplistix.co.uk > -- http://www.voidspace.org.uk/ 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... URL: From fuzzyman at gmail.com Wed Apr 6 19:32:56 2011 From: fuzzyman at gmail.com (Michael Foord) Date: Wed, 6 Apr 2011 18:32:56 +0100 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> Message-ID: On 6 April 2011 18:24, Tres Seaver wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On 04/06/2011 12:52 PM, Michael Foord wrote: > > On 6 April 2011 17:44, P.J. Eby wrote: > > > >> At 02:13 AM 4/6/2011 -0400, Terry Reedy wrote: > >> > >>> No, he countered with a community poll that has, as I remember, 100-200 > >>> responders. After a contentious discussion. > >>> > >> > >> A poll which, unfortunately, had numerous choices about what *kind* of > >> system to have, such that votes *against* Martin's proposal were widely > >> split, and IIRC, the difference between the number of votes cast for > >> different alternatives was pretty negligible. > >> > >> Looking at the numbers at one point, I concluded that if I had run a > poll > >> with those results, I would have had to conclude that there was > essentially > >> zero consensus about what direction should be taken, and either tabled > the > >> notion or gone back to the list to try to get more specific discussion. > >> > >> Unfortunately, the poll was run as a decision-making mechanism, rather > than > >> an information-gathering one. > >> > >> > > > > Well sure, but it still didn't show that *nobody* wants ratings which is > > what people in this thread seem to be claiming. > > > > I'm afraid that those in this thread, and with all respect to Jacob those > > who respond to him on twitter, *do* represent a vocal minority primarily > of > > package authors and are not in any way representative of users of pypi. A > > poll may be flawed however it is done, but is a much better mechanism. > > > > Martin *has* offered to do another poll, an offer that has been ignored. > > For those who may not be familiar with the prior poll, here are the > results reported to the list[1]: > > - ----------------------------------- %< -------------------------------- > Allow ratings and comments on all packages (status quo) 237 > Allow package owners to disallow comments (ratings unmodified). 139 > Allow comments, but only send them to package owners (ratings > unmodified). 33 > Disallow comments (ratings unmodified). 24 > Disallow ratings and comments (status three months ago). 88 > - ----------------------------------- %< -------------------------------- > > Hmmm... despite claims that this poll was flawed it seems the result is pretty clear. A *big* majority of users who voted were in favour of ratings then. But I guess that is another argument. *sigh* Michael > Interpretation of those results was the subject of a huge thread, which > produced no clear consensus (at least to me). > > > [1] http://permalink.gmane.org/gmane.comp.python.catalog/2169 > > > Tres. > - -- > =================================================================== > Tres Seaver +1 540-429-0999 tseaver at palladion.com > Palladion Software "Excellence by Design" http://palladion.com > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.10 (GNU/Linux) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ > > iEYEARECAAYFAk2cocQACgkQ+gerLs4ltQ5rcQCgqA1G5mw21Zvz++iMM3vc63i0 > McwAoLz3aKjZ5bEg07xAcBwhHC8irMjz > =caM8 > -----END PGP SIGNATURE----- > > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- http://www.voidspace.org.uk/ 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... URL: From pje at telecommunity.com Wed Apr 6 19:43:35 2011 From: pje at telecommunity.com (P.J. Eby) Date: Wed, 06 Apr 2011 13:43:35 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> Message-ID: <20110406174352.A94F03A4077@sparrow.telecommunity.com> At 05:52 PM 4/6/2011 +0100, Michael Foord wrote: >On 6 April 2011 17:44, P.J. Eby ><pje at telecommunity.com> wrote: >Unfortunately, the poll was run as a decision-making mechanism, >rather than an information-gathering one. > >Well sure, but it still didn't show that *nobody* wants ratings >which is what people in this thread seem to be claiming. But it *did* show that the vast majority of people did NOT want what we ended up with. >Martin *has* offered to do another poll, an offer that has been ignored. Without some discussion of what that poll is supposed to accomplish, and a sensible design of the poll, it would be pointless. An information-gathering survey might be useful, asking people to indicate their use cases and reasoning for or against a commenting system, as well as asking whether they were aware that there is already a rating/comment system. But that's an entirely different thing from putting to a vote what features a rating system should have and using it to make a decision. (Which is what the previous "poll" was.) From pje at telecommunity.com Wed Apr 6 20:03:14 2011 From: pje at telecommunity.com (P.J. Eby) Date: Wed, 06 Apr 2011 14:03:14 -0400 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> Message-ID: <20110406180331.EE4C93A4077@sparrow.telecommunity.com> At 06:32 PM 4/6/2011 +0100, Michael Foord wrote: >Hmmm... despite claims that this poll was flawed it seems the result >is pretty clear. A *big* majority of users who voted were in favour >of ratings then. Do note that my interpretation was not about *ratings*; it was that the poll showed a majority being against *mandatory public comments*. (And it does, 284 to 237 against, although in all honesty, it seems the numbers I was recalling were ones that I saw mid-way through the poll, rather than at its end. So, the majority is not by as large of a margin as I was recalling.) From martin at v.loewis.de Wed Apr 6 20:13:52 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 06 Apr 2011 20:13:52 +0200 Subject: [Catalog-sig] Please turn off ratings In-Reply-To: References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> Message-ID: <4D9CAD60.1020409@v.loewis.de> > It seems to me that, given we're quite a long way into the potential > user adoption of this feature, it's simply not being used. I would > support removing it, as it seems that the only purpose it serves is to > antagonise. Thanks for your advise. I (now) believe this is a reasonable action to take, even though I really dislike not being able to defend what I believe is in the interest of the PyPI users. Kind regards, Martin From jjl at pobox.com Wed Apr 6 21:22:44 2011 From: jjl at pobox.com (John J Lee) Date: Wed, 6 Apr 2011 20:22:44 +0100 (BST) Subject: [Catalog-sig] Please turn off ratings In-Reply-To: <4D9CAD60.1020409@v.loewis.de> References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> <4D9CAD60.1020409@v.loewis.de> Message-ID: On Wed, 6 Apr 2011, "Martin v. L?wis" wrote: >> It seems to me that, given we're quite a long way into the potential >> user adoption of this feature, it's simply not being used. I would >> support removing it, as it seems that the only purpose it serves is to >> antagonise. > > Thanks for your advise. I (now) believe this is a reasonable action > to take, even though I really dislike not being able to defend what > I believe is in the interest of the PyPI users. Before this thread gets even longer: project release pages on the pypi.python.org website now say: """ The rating feature has been removed. See catalog-sig for the discussion of this removal. """ John From monitor at jacobian.org Thu Apr 7 01:54:19 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 18:54:19 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection failed Message-ID: <1302134065.1@jacobian.org> Connection failed Service pypi.python.org Date: Wed, 06 Apr 2011 18:54:19 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Thu Apr 7 03:30:18 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Wed, 06 Apr 2011 20:30:18 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection succeeded Message-ID: <1302139822.1@jacobian.org> Connection succeeded Service pypi.python.org Date: Wed, 06 Apr 2011 20:30:18 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[pypi.python.org:80] via TCP Your faithful employee, monit From ben+python at benfinney.id.au Thu Apr 7 07:28:10 2011 From: ben+python at benfinney.id.au (Ben Finney) Date: Thu, 07 Apr 2011 15:28:10 +1000 Subject: [Catalog-sig] Please turn off ratings References: <4D9B73B1.2050804@v.loewis.de> <4D9B7EB7.3070207@v.loewis.de> <20110405230319.08A133A4063@sparrow.telecommunity.com> <4D9BA0F7.1060500@v.loewis.de> <4D9BAD36.3010002@v.loewis.de> <20110406164501.310DC3A4077@sparrow.telecommunity.com> <4D9CA1BD.30406@simplistix.co.uk> <4D9CA28F.4030907@simplistix.co.uk> Message-ID: <87pqoyfx5h.fsf@benfinney.id.au> Chris Withers writes: > On 06/04/2011 18:25, Michael Foord wrote: > > Would it be fair to split the yes vote by offering alternatives? [?] > It just dilutes what we're trying to establish here: do people want > ratings on PyPI? Yes or no? That doesn't address the ?wishing for ponies? problem Laura Creighton raised. From a ?yes? answer to that question, you have no information about what the respondent thinks ?ratings on PyPI? actually entails, nor what they're unwilling to compromise to have it happen. ?Do you want better public education?? is far more likely than ?Do you want funding diverted from health care to education?? to receive ?yes? answers. The information from the former poll doesn't capture much useful. -- \ ?No matter how cynical you become, it's never enough to keep | `\ up.? ?Jane Wagner, via Lily Tomlin | _o__) | Ben Finney From ben+python at benfinney.id.au Thu Apr 7 07:30:30 2011 From: ben+python at benfinney.id.au (Ben Finney) Date: Thu, 07 Apr 2011 15:30:30 +1000 Subject: [Catalog-sig] Please turn off ratings References: <4D9B73B1.2050804@v.loewis.de> <4D9B834B.4070109@v.loewis.de> <4D9B8CC5.3040701@v.loewis.de> <4D9CAD60.1020409@v.loewis.de> Message-ID: <87lizmfx1l.fsf@benfinney.id.au> "Martin v. L?wis" writes: > > It seems to me that, given we're quite a long way into the potential > > user adoption of this feature, it's simply not being used. I would > > support removing it, as it seems that the only purpose it serves is > > to antagonise. > > Thanks for your advise. I (now) believe this is a reasonable action > to take, even though I really dislike not being able to defend what > I believe is in the interest of the PyPI users. Thank you for that decision, and for being honest about the difficulty in making it. -- \ ?? whoever claims any right that he is unwilling to accord to | `\ his fellow-men is dishonest and infamous.? ?Robert G. | _o__) Ingersoll, _The Liberty of Man, Woman and Child_, 1877 | Ben Finney From techtonik at gmail.com Thu Apr 7 07:41:19 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Thu, 7 Apr 2011 08:41:19 +0300 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) Message-ID: On Thu, Apr 7, 2011 at 8:25 AM, python.org wiki wrote: > > ?You can access JSON information about packages by using the URL format > > ? ?http://pypi.python.org/pypi/package_name/json Why Content-Disposition: attachment; ??? Please, CC. >python -m wget http://pypi.python.org/pypi/wget/json 8192 / unknown Date: Thu, 07 Apr 2011 05:34:16 GMT Server: Apache/2.2.16 (Debian) Content-Disposition: attachment; filename=wget-0.7.json Connection: close Content-Type: application/json; charset="UTF-8" Saved under json -- anatoly t. From richard at python.org Thu Apr 7 07:52:13 2011 From: richard at python.org (Richard Jones) Date: Thu, 7 Apr 2011 15:52:13 +1000 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: References: Message-ID: On Thu, Apr 7, 2011 at 3:41 PM, anatoly techtonik wrote: > On Thu, Apr 7, 2011 at 8:25 AM, python.org wiki wrote: >> >> ?You can access JSON information about packages by using the URL format >> >> ? ?http://pypi.python.org/pypi/package_name/json > > Why Content-Disposition: attachment; ??? > Please, CC. Why not? IIRC when I implemented it at the EuroPython sprints I asked around those near me and the consensus was for attachment. It should have no bearing on programmatic use, and I find it more convenient for in-browser use (since application/json is not handled). >>python -m wget http://pypi.python.org/pypi/wget/json > 8192 / unknown > Date: Thu, 07 Apr 2011 05:34:16 GMT > Server: Apache/2.2.16 (Debian) > Content-Disposition: attachment; filename=wget-0.7.json > Connection: close > Content-Type: application/json; charset="UTF-8" > > Saved under json > > -- > anatoly t. > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > > From techtonik at gmail.com Thu Apr 7 08:05:48 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Thu, 7 Apr 2011 09:05:48 +0300 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: References: Message-ID: On Thu, Apr 7, 2011 at 8:52 AM, Richard Jones wrote: > On Thu, Apr 7, 2011 at 3:41 PM, anatoly techtonik wrote: >> On Thu, Apr 7, 2011 at 8:25 AM, python.org wiki wrote: >>> >>> ?You can access JSON information about packages by using the URL format >>> >>> ? ?http://pypi.python.org/pypi/package_name/json >> >> Why Content-Disposition: attachment; ??? >> Please, CC. > > Why not? IIRC when I implemented it at the EuroPython sprints I asked > around those near me and the consensus was for attachment. It should > have no bearing on programmatic use, and I find it more convenient for > in-browser use (since application/json is not handled). What do you mean by in-browser use? Chrome downloads this URL instead of displaying contents inline, and it becomes much harder to debug. Please, CC. -- anatoly t. From richard at python.org Thu Apr 7 08:19:11 2011 From: richard at python.org (Richard Jones) Date: Thu, 7 Apr 2011 16:19:11 +1000 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: References: Message-ID: On Thu, Apr 7, 2011 at 4:05 PM, anatoly techtonik wrote: > On Thu, Apr 7, 2011 at 8:52 AM, Richard Jones wrote: >> On Thu, Apr 7, 2011 at 3:41 PM, anatoly techtonik wrote: >>> On Thu, Apr 7, 2011 at 8:25 AM, python.org wiki wrote: >>>> >>>> ?You can access JSON information about packages by using the URL format >>>> >>>> ? ?http://pypi.python.org/pypi/package_name/json >>> >>> Why Content-Disposition: attachment; ??? >>> Please, CC. >> >> Why not? IIRC when I implemented it at the EuroPython sprints I asked >> around those near me and the consensus was for attachment. It should >> have no bearing on programmatic use, and I find it more convenient for >> in-browser use (since application/json is not handled). > > What do you mean by in-browser use? Chrome downloads this URL instead > of displaying contents inline, and it becomes much harder to debug. That is in-browser use. When I was testing the browser in use at the time (firefox? chrome?) would complain that there was no application assigned to application/json content. It seemed a better solution to have the file download to disk where the user could choose what to do with it. And as I said, the people present at the sprint at the time agreed :-) I've just turned off that header, and it appears Chromium nightly (from a couple of days ago) displays the content in the browser with no fuss. The latest Firefox still complains, but the latest Safari is ok with just dumping it in the page. Does anyone have any objections to removing the "Content-Disposition: attachment"? Richard From lac at openend.se Thu Apr 7 10:58:58 2011 From: lac at openend.se (Laura Creighton) Date: Thu, 07 Apr 2011 10:58:58 +0200 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: Message from Richard Jones of "Thu, 07 Apr 2011 16:19:11 +1000." References: Message-ID: <201104070858.p378wwxb025144@theraft.openend.se> In a message of Thu, 07 Apr 2011 16:19:11 +1000, Richard Jones writes: >I've just turned off that header, and it appears Chromium nightly >(from a couple of days ago) displays the content in the browser with >no fuss. The latest Firefox still complains, but the latest Safari is >ok with just dumping it in the page. > >Does anyone have any objections to removing the "Content-Disposition: >attachment"? > > > Richard If you use this FF extension https://addons.mozilla.org/en-us/firefox/addon/open-in-browser/ (which I think should be part of default FF, but nevermind) does the problem go away? Laura From pydanny at gmail.com Thu Apr 7 18:56:20 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Thu, 7 Apr 2011 09:56:20 -0700 Subject: [Catalog-sig] Can the 'score' system be explained on PyPI? Message-ID: In doing research on Python powered SOAP libraries I was advised to check out SUDS. My PyPI search (http://pypi.python.org/pypi?%3Aaction=search&term=SOAP&submit=search) showed that SUDS has a '2' score. Nowhere on the search results page is 'score' explained. Is it how likely it turns up in searches? The number of downloads compared to everything else? Or... dare I say... the results of a rating system? Fixing this one is easy. Just put some text at the top of the search page explaining what 'score' means. Thanks! -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From bob at redivi.com Thu Apr 7 19:15:55 2011 From: bob at redivi.com (Bob Ippolito) Date: Thu, 7 Apr 2011 10:15:55 -0700 Subject: [Catalog-sig] Can the 'score' system be explained on PyPI? In-Reply-To: References: Message-ID: On Thu, Apr 7, 2011 at 9:56 AM, Daniel Greenfeld wrote: > In doing research on Python powered SOAP libraries I was advised to > check out SUDS. My PyPI search > (http://pypi.python.org/pypi?%3Aaction=search&term=SOAP&submit=search) > showed that SUDS has a '2' score. > > Nowhere on the search results page is 'score' explained. Is it how > likely it turns up in searches? The number of downloads compared to > everything else? Or... dare I say... the results of a rating system? > > Fixing this one is easy. Just put some text at the top of the search > page explaining what 'score' means. There's a title tag on that element, so in most browsers if you let your cursor hover over it you'll see the following: Occurrence of search term weighted by field (name, summary, keywords, description, author, maintainer) -bob From renesd at gmail.com Thu Apr 7 19:17:05 2011 From: renesd at gmail.com (=?ISO-8859-1?Q?Ren=E9_Dudfield?=) Date: Thu, 7 Apr 2011 18:17:05 +0100 Subject: [Catalog-sig] Can the 'score' system be explained on PyPI? In-Reply-To: References: Message-ID: Hey hey. That's the relevance score. Or how relevant pypi thinks the results are to your search query. It could easily be mistaken for other meanings though as you suggest. It's kind of common on some websites search features. Another fix could be: s/Score/Relevance/g cya! On Thu, Apr 7, 2011 at 5:56 PM, Daniel Greenfeld wrote: > In doing research on Python powered SOAP libraries I was advised to > check out SUDS. My PyPI search > (http://pypi.python.org/pypi?%3Aaction=search&term=SOAP&submit=search) > showed that SUDS has a '2' score. > > Nowhere on the search results page is 'score' explained. Is it how > likely it turns up in searches? The number of downloads compared to > everything else? Or... dare I say... the results of a rating system? > > Fixing this one is easy. Just put some text at the top of the search > page explaining what 'score' means. > > Thanks! > > -- > 'Knowledge is Power' > Daniel Greenfeld > http://pydanny.com > http://cartwheelweb.com > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -------------- next part -------------- An HTML attachment was scrubbed... URL: From richard at python.org Fri Apr 8 02:56:24 2011 From: richard at python.org (Richard Jones) Date: Fri, 8 Apr 2011 10:56:24 +1000 Subject: [Catalog-sig] Can the 'score' system be explained on PyPI? In-Reply-To: References: Message-ID: On Fri, Apr 8, 2011 at 3:17 AM, Ren? Dudfield wrote: > That's the relevance score.? Or how relevant pypi thinks the results are to > your search query.? It could easily be mistaken for other meanings though as > you suggest.? It's kind of common on some websites search features. > > Another fix could be: s/Score/Relevance/g I have chosen the even less "score"-like term "weight" and copied the explanation text to the footer of the page. Richard From richard at python.org Fri Apr 8 04:18:51 2011 From: richard at python.org (Richard Jones) Date: Fri, 8 Apr 2011 12:18:51 +1000 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: <201104070858.p378wwxb025144@theraft.openend.se> References: <201104070858.p378wwxb025144@theraft.openend.se> Message-ID: On Thu, Apr 7, 2011 at 6:58 PM, Laura Creighton wrote: > In a message of Thu, 07 Apr 2011 16:19:11 +1000, Richard Jones writes: > > > >>I've just turned off that header, and it appears Chromium nightly >>(from a couple of days ago) displays the content in the browser with >>no fuss. The latest Firefox still complains, but the latest Safari is >>ok with just dumping it in the page. >> >>Does anyone have any objections to removing the "Content-Disposition: >>attachment"? > > If you use this FF extension > https://addons.mozilla.org/en-us/firefox/addon/open-in-browser/ > (which I think should be part of default FF, but nevermind) does the > problem go away? Interesting - but I'm not sure that's a great solution. Can someone point IE at http://pypi.python.org/pypi/roundup/json and tell me what it does please? Richard From lac at openend.se Fri Apr 8 05:34:29 2011 From: lac at openend.se (Laura Creighton) Date: Fri, 08 Apr 2011 05:34:29 +0200 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: Message from Richard Jones of "Fri, 08 Apr 2011 12:18:51 +1000." References: <201104070858.p378wwxb025144@theraft.openend.se> Message-ID: <201104080334.p383YTmc028273@theraft.openend.se> In a message of Fri, 08 Apr 2011 12:18:51 +1000, Richard Jones writes: >>>Does anyone have any objections to removing the "Content-Disposition: >>>attachment"? >> >> If you use this FF extension >> https://addons.mozilla.org/en-us/firefox/addon/open-in-browser/ >> (which I think should be part of default FF, but nevermind) does the >> problem go away? > >Interesting - but I'm not sure that's a great solution. Me either, but it seemed germane to the discussion, at any rate. >Can someone point IE at http://pypi.python.org/pypi/roundup/json and >tell me what it does please? Let me know if you don't get a response in 3 or 4 hours and I will try it when I am at work. > Richard Laura From techtonik at gmail.com Fri Apr 8 10:24:56 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Fri, 8 Apr 2011 11:24:56 +0300 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: References: <201104070858.p378wwxb025144@theraft.openend.se> Message-ID: On Fri, Apr 8, 2011 at 5:18 AM, Richard Jones wrote: > > Interesting - but I'm not sure that's a great solution. > > Can someone point IE at http://pypi.python.org/pypi/roundup/json and > tell me what it does please? IE7 proposes download of 'json' file. -- anatoly t. From richard at python.org Fri Apr 8 10:31:47 2011 From: richard at python.org (Richard Jones) Date: Fri, 8 Apr 2011 18:31:47 +1000 Subject: [Catalog-sig] JSON content-disposition (Was: [PythonInfo Wiki] Update of "CheeseShop" by RichardJones) In-Reply-To: References: <201104070858.p378wwxb025144@theraft.openend.se> Message-ID: On Fri, Apr 8, 2011 at 6:24 PM, anatoly techtonik wrote: > On Fri, Apr 8, 2011 at 5:18 AM, Richard Jones wrote: >> Interesting - but I'm not sure that's a great solution. >> >> Can someone point IE at http://pypi.python.org/pypi/roundup/json and >> tell me what it does please? > > IE7 proposes download of 'json' file. As of about two hours ago I switched the code over to not using "Content-Disposition: attachment". Based on the current browser behaviour I figure it's appropriate. Thanks for bearing with me while I investigated, anatoly. Richard From pydanny at gmail.com Fri Apr 8 19:01:19 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Fri, 8 Apr 2011 10:01:19 -0700 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? Message-ID: Projects have home pages, they also have Repos but that is not on PyPI. For the forthcoming Python Packages (https://github.com/cartwheelweb/packaginator) having this field would mean gathering metrics on those projects would be great. Otherwise we have to check 10K+ records. I know that the data will be entered in eventually, but eventually we'll have the canonical repo source for a greater majority of projects.. -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From ziade.tarek at gmail.com Fri Apr 8 19:09:23 2011 From: ziade.tarek at gmail.com (=?ISO-8859-1?Q?Tarek_Ziad=E9?=) Date: Fri, 8 Apr 2011 10:09:23 -0700 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: I added in PEP 345 a field where you can store any number of urls for the project Check it out, PyPI also support displaying it under a "Project Links" portlet example: http://pypi.python.org/pypi/gfbi_core/0.2 On Fri, Apr 8, 2011 at 10:01 AM, Daniel Greenfeld wrote: > Projects have home pages, they also have Repos but that is not on > PyPI. For the forthcoming Python Packages > (https://github.com/cartwheelweb/packaginator) having this field would > mean gathering metrics on those projects would be great. Otherwise we > have to check 10K+ records. I know that the data will be entered in > eventually, but eventually we'll have the canonical repo source for a > greater majority of projects.. > > -- > 'Knowledge is Power' > Daniel Greenfeld > http://pydanny.com > http://cartwheelweb.com > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- Tarek Ziad? | http://ziade.org From pydanny at gmail.com Fri Apr 8 19:24:49 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Fri, 8 Apr 2011 10:24:49 -0700 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: Hey Tarek, In PEP-345 there isn't a specific URL field. Instead we have to hope that people add something and then follow a 'source code' label convention. To know to add this field, one needs to know the convention, and that is in PEP-345, far from PyPI and its instructions. Even if instructed, from a sociological point of view one could argue that this is not the most efficient way to handle the issue. My request is for a obvious, specific field. An obvious, specific field will encourage package owners to provide links to their source code repositories. The field and the APIs it will feed will allow tools like Packaginator, Read the Docs, Depot.io, and other yet-to-be-invented tools a clear path to the source code repo for existing and unknown reasons. Thanks! Danny Greenfeld On Fri, Apr 8, 2011 at 10:09 AM, Tarek Ziad? wrote: > I added in PEP 345 a field where you can store any number of urls for > the project > > Check it out, PyPI also support displaying it under a "Project Links" portlet > > example: http://pypi.python.org/pypi/gfbi_core/0.2 > > > > On Fri, Apr 8, 2011 at 10:01 AM, Daniel Greenfeld wrote: >> Projects have home pages, they also have Repos but that is not on >> PyPI. For the forthcoming Python Packages >> (https://github.com/cartwheelweb/packaginator) having this field would >> mean gathering metrics on those projects would be great. Otherwise we >> have to check 10K+ records. I know that the data will be entered in >> eventually, but eventually we'll have the canonical repo source for a >> greater majority of projects.. >> >> -- >> 'Knowledge is Power' >> Daniel Greenfeld >> http://pydanny.com >> http://cartwheelweb.com >> _______________________________________________ >> Catalog-SIG mailing list >> Catalog-SIG at python.org >> http://mail.python.org/mailman/listinfo/catalog-sig >> > > > > -- > Tarek Ziad? | http://ziade.org > -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From ziade.tarek at gmail.com Fri Apr 8 19:32:48 2011 From: ziade.tarek at gmail.com (=?ISO-8859-1?Q?Tarek_Ziad=E9?=) Date: Fri, 8 Apr 2011 10:32:48 -0700 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: On Fri, Apr 8, 2011 at 10:24 AM, Daniel Greenfeld wrote: > Hey Tarek, > > In PEP-345 there isn't a specific URL field. Instead we have to hope > that people add something and then follow a 'source code' label > convention. To know to add this field, one needs to know the > convention, and that is in PEP-345, far from PyPI and its > instructions. Even if instructed, from a sociological point of view > one could argue that this is not the most efficient way to handle the > issue. > > My request is for a obvious, specific field. An obvious, specific > field will encourage package owners to provide links to their source > code repositories. The field and the APIs it will feed will allow > tools like Packaginator, Read the Docs, Depot.io, and other > yet-to-be-invented tools a clear path to the source code repo for > existing and unknown reasons. We had that discussion in the past when we added the Project-URL field. We ended up with the conclusion that it's hard to define such a field in a generic way, so it's free-form for now. Mostly because of the url schemes, and because we were not sure what people would do with it. What do you want to do ? do you want a web UI of the repo ? or a git://, ssh:// lp:// link ? etc.. what is expected in the linked page ? Do you want a single repo link ? several ? what if the project does not have a repo ? Cheers > > Danny Greenfeld > > On Fri, Apr 8, 2011 at 10:09 AM, Tarek Ziad? wrote: >> I added in PEP 345 a field where you can store any number of urls for >> the project >> >> Check it out, PyPI also support displaying it under a "Project Links" portlet >> >> example: http://pypi.python.org/pypi/gfbi_core/0.2 >> >> >> >> On Fri, Apr 8, 2011 at 10:01 AM, Daniel Greenfeld wrote: >>> Projects have home pages, they also have Repos but that is not on >>> PyPI. For the forthcoming Python Packages >>> (https://github.com/cartwheelweb/packaginator) having this field would >>> mean gathering metrics on those projects would be great. Otherwise we >>> have to check 10K+ records. I know that the data will be entered in >>> eventually, but eventually we'll have the canonical repo source for a >>> greater majority of projects.. >>> >>> -- >>> 'Knowledge is Power' >>> Daniel Greenfeld >>> http://pydanny.com >>> http://cartwheelweb.com >>> _______________________________________________ >>> Catalog-SIG mailing list >>> Catalog-SIG at python.org >>> http://mail.python.org/mailman/listinfo/catalog-sig >>> >> >> >> >> -- >> Tarek Ziad? | http://ziade.org >> > > > > -- > 'Knowledge is Power' > Daniel Greenfeld > http://pydanny.com > http://cartwheelweb.com > -- Tarek Ziad? | http://ziade.org From exarkun at twistedmatrix.com Fri Apr 8 19:58:37 2011 From: exarkun at twistedmatrix.com (exarkun at twistedmatrix.com) Date: Fri, 08 Apr 2011 17:58:37 -0000 Subject: [Catalog-sig] pypissh Message-ID: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> Hello, I tried pypissh 1.1. When I ran it with no arguments, it produced this result: Traceback (most recent call last): File "pypissh.py", line 85, in f = urlopen('httpssh://submit at pypi.python.org/pypi') File "pypissh.py", line 74, in urlopen return _opener.open(req, data, timeout) File "pypissh.py", line 57, in open if req.get_full_url().startswith(_badprefix): AttributeError: 'str' object has no attribute 'get_full_url' Jean-Paul From exarkun at twistedmatrix.com Fri Apr 8 20:18:49 2011 From: exarkun at twistedmatrix.com (exarkun at twistedmatrix.com) Date: Fri, 08 Apr 2011 18:18:49 -0000 Subject: [Catalog-sig] pypissh In-Reply-To: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> References: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> Message-ID: <20110408181849.1992.529337496.divmod.xquotient.361@localhost.localdomain> On 05:58 pm, exarkun at twistedmatrix.com wrote: >Hello, > >I tried pypissh 1.1. When I ran it with no arguments, it produced this >result: > >Traceback (most recent call last): > File "pypissh.py", line 85, in > f = urlopen('httpssh://submit at pypi.python.org/pypi') > File "pypissh.py", line 74, in urlopen > return _opener.open(req, data, timeout) > File "pypissh.py", line 57, in open > if req.get_full_url().startswith(_badprefix): >AttributeError: 'str' object has no attribute 'get_full_url' Beyond this, if I try to use pypissh with a simple setup.py like this: import pypissh pypissh.monkeypatch() from distutils.core import setup setup( name="fake-experiment", version="1") It fails with this output: running sdist warning: sdist: missing required meta-data: url warning: sdist: missing meta-data: either (author and author_email) or (maintainer and maintainer_email) must be supplied reading manifest file 'MANIFEST' creating fake-experiment-1 making hard links in fake-experiment-1... hard linking README -> fake-experiment-1 hard linking setup.py -> fake-experiment-1 tar -cf dist/fake-experiment-1.tar fake-experiment-1 gzip -f9 dist/fake-experiment-1.tar removing 'fake-experiment-1' (and everything under it) running upload Submitting dist/fake-experiment-1.tar.gz to http://submit at pypi.python.org/pypi [Errno -2] Name or service not known >Jean-Paul >_______________________________________________ >Catalog-SIG mailing list >Catalog-SIG at python.org >http://mail.python.org/mailman/listinfo/catalog-sig From martin at v.loewis.de Fri Apr 8 21:08:52 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Fri, 08 Apr 2011 21:08:52 +0200 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: <4D9F5D44.1030703@v.loewis.de> > In PEP-345 there isn't a specific URL field. Instead we have to hope > that people add something and then follow a 'source code' label > convention. [...] > My request is for a obvious, specific field. I don't think these need to contradict each other: it could be a Project-URL, but still could be an obvious, specific field. Furthermore, I could envision PyPI setting standard labels. I wouldn't call it "code repository URL", but "Subversion", "Mercurial", or "git", with a free-form option of putting in an arbitrary label. Regards, Martin From jacob at jacobian.org Fri Apr 8 21:30:10 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Fri, 8 Apr 2011 14:30:10 -0500 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: On Fri, Apr 8, 2011 at 12:09 PM, Tarek Ziad? wrote: > I added in PEP 345 a field where you can store any number of urls for > the project > > Check it out, PyPI also support displaying it under a "Project Links" portlet > > example: http://pypi.python.org/pypi/gfbi_core/0.2 It looks like these don't show up on the /simple pages, making scraping of those links a bit complicated and brittle. Can that be added? Or is there another way to fetch these links more programmatically? Danny: my suggestion, given that PEP 345's already implemented, is to just pick a Project-URL label, document it, and encourage people to use it for source control. Python Packages is the only thing that needs it right now, so you can just paint the bikeshed however you want and establish a de facto standard. Jacob From martin at v.loewis.de Fri Apr 8 22:07:03 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Fri, 08 Apr 2011 22:07:03 +0200 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: <4D9F6AE7.40303@v.loewis.de> Am 08.04.2011 21:30, schrieb Jacob Kaplan-Moss: > On Fri, Apr 8, 2011 at 12:09 PM, Tarek Ziad? wrote: >> I added in PEP 345 a field where you can store any number of urls for >> the project >> >> Check it out, PyPI also support displaying it under a "Project Links" portlet >> >> example: http://pypi.python.org/pypi/gfbi_core/0.2 > > It looks like these don't show up on the /simple pages, making > scraping of those links a bit complicated and brittle. Can that be > added? Or is there another way to fetch these links more > programmatically? It's available through XML-RPC: py> p=xmlrpclib.ServerProxy("http://pypi.python.org/pypi") py> p.release_data('Distutils2','1.0a1')['project_url'] ['Repository,http://hg.python.org/distutils2', 'Mailing-list,http://mail.python.org/mailman/listinfo/distutils-sig/', 'Documentation,http://packages.python.org/Distutils2', 'Bug tracker,http://bugs.python.org'] I wouldn't put it on the simple page, since it's not related to installing the software. HTH, Martin From tseaver at palladion.com Sat Apr 9 00:57:50 2011 From: tseaver at palladion.com (Tres Seaver) Date: Fri, 08 Apr 2011 18:57:50 -0400 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: <4D9F6AE7.40303@v.loewis.de> References: <4D9F6AE7.40303@v.loewis.de> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/08/2011 04:07 PM, "Martin v. L?wis" wrote: > Am 08.04.2011 21:30, schrieb Jacob Kaplan-Moss: >> On Fri, Apr 8, 2011 at 12:09 PM, Tarek Ziad? wrote: >>> I added in PEP 345 a field where you can store any number of urls for >>> the project >>> >>> Check it out, PyPI also support displaying it under a "Project Links" portlet >>> >>> example: http://pypi.python.org/pypi/gfbi_core/0.2 >> >> It looks like these don't show up on the /simple pages, making >> scraping of those links a bit complicated and brittle. Can that be >> added? Or is there another way to fetch these links more >> programmatically? > > It's available through XML-RPC: > > py> p=xmlrpclib.ServerProxy("http://pypi.python.org/pypi") > py> p.release_data('Distutils2','1.0a1')['project_url'] > ['Repository,http://hg.python.org/distutils2', > 'Mailing-list,http://mail.python.org/mailman/listinfo/distutils-sig/', > 'Documentation,http://packages.python.org/Distutils2', 'Bug > tracker,http://bugs.python.org'] > > I wouldn't put it on the simple page, since it's not related to > installing the software. I nodded to myself in agreement at first, and then remembered that setuptools is willing to download and install from at least Subversion URLs: $ /opt/Python-2.7.0/bin/virtualenv --no-site-packages \ /tmp/foo ... $ env | grep RSVN RSVN=svn+ssh://repoze at svn.repoze.org/svn $ /tmp/foo/bin/easy_install $RSVN/compoze/trunk Doing subversion checkout from svn+ssh://repoze at svn.repoze.org\ /svn/compoze/trunk to /tmp/easy_install-Giyqv5/trunk ... Finished processing dependencies for compoze==0.4 I don't know if it will do so from Mercurial, Git, or Bazaar URLs (maybe if you have the appropriate plugins installed?) I also don't know whether it scrapes VCS URLs from the '/simple' interface. Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver at palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk2fku4ACgkQ+gerLs4ltQ5APQCfRnxaznbzEJc9ht8bsCIGAQ2A uqoAn2B3wOIbTG+8NTbTZZgqj8a/+X+L =yMku -----END PGP SIGNATURE----- From richard at python.org Sat Apr 9 01:13:19 2011 From: richard at python.org (Richard Jones) Date: Sat, 9 Apr 2011 09:13:19 +1000 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: <4D9F6AE7.40303@v.loewis.de> Message-ID: On Sat, Apr 9, 2011 at 8:57 AM, Tres Seaver wrote: >I also don't know > whether it [setuptools] scrapes VCS URLs from the '/simple' interface. Oh gods I hope not. Richard From pydanny at gmail.com Sat Apr 9 20:43:14 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Sat, 9 Apr 2011 11:43:14 -0700 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: Message-ID: On Fri, Apr 8, 2011 at 12:30 PM, Jacob Kaplan-Moss wrote: > On Fri, Apr 8, 2011 at 12:09 PM, Tarek Ziad? wrote: > Danny: my suggestion, given that PEP 345's already implemented, is to > just pick a Project-URL label, document it, and encourage people to > use it for source control. Python Packages is the only thing that > needs it right now, so you can just paint the bikeshed however you > want and establish a de facto standard. Thanks for the suggestion Jacob. The bikeshed will be painted quite soon. :) -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From pje at telecommunity.com Sat Apr 9 21:12:00 2011 From: pje at telecommunity.com (P.J. Eby) Date: Sat, 09 Apr 2011 15:12:00 -0400 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: <4D9F6AE7.40303@v.loewis.de> Message-ID: <20110409191155.B3C933A4077@sparrow.telecommunity.com> At 09:13 AM 4/9/2011 +1000, Richard Jones wrote: >On Sat, Apr 9, 2011 at 8:57 AM, Tres Seaver wrote: > >I also don't know > > whether it [setuptools] scrapes VCS URLs from the '/simple' interface. > >Oh gods I hope not. Um, that's a big part of what the /simple interface is *for*, so it doesn't have to pull them from the formatted HTML of the /pypi interface. (As for non-SVN revision control systems, right now it does not support their special URLs, even with a plugin.) From richard at python.org Sun Apr 10 00:16:39 2011 From: richard at python.org (Richard Jones) Date: Sun, 10 Apr 2011 08:16:39 +1000 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: <20110409191155.B3C933A4077@sparrow.telecommunity.com> References: <4D9F6AE7.40303@v.loewis.de> <20110409191155.B3C933A4077@sparrow.telecommunity.com> Message-ID: On Sun, Apr 10, 2011 at 5:12 AM, P.J. Eby wrote: > At 09:13 AM 4/9/2011 +1000, Richard Jones wrote: >> On Sat, Apr 9, 2011 at 8:57 AM, Tres Seaver wrote: >> >I also don't know >> > whether it [setuptools] scrapes VCS URLs from the '/simple' interface. >> >> Oh gods I hope not. > > Um, that's a big part of what the /simple interface is *for*, so it doesn't > have to pull them from the formatted HTML of the /pypi interface. Just to be clear: the VCS URL is only ever used in the absence of any other (suitable) package files? That's my reading of http://peak.telecommunity.com/DevCenter/EasyInstall#downloading-and-installing-a-package Richard From pje at telecommunity.com Sun Apr 10 00:37:03 2011 From: pje at telecommunity.com (P.J. Eby) Date: Sat, 09 Apr 2011 18:37:03 -0400 Subject: [Catalog-sig] Can PyPI have a new field that stores the code repository URL? In-Reply-To: References: <4D9F6AE7.40303@v.loewis.de> <20110409191155.B3C933A4077@sparrow.telecommunity.com> Message-ID: <20110409223705.2C1CE3A4077@sparrow.telecommunity.com> At 08:16 AM 4/10/2011 +1000, Richard Jones wrote: >On Sun, Apr 10, 2011 at 5:12 AM, P.J. Eby wrote: > > At 09:13 AM 4/9/2011 +1000, Richard Jones wrote: > >> On Sat, Apr 9, 2011 at 8:57 AM, Tres Seaver wrote: > >> >I also don't know > >> > whether it [setuptools] scrapes VCS URLs from the '/simple' interface. > >> > >> Oh gods I hope not. > > > > Um, that's a big part of what the /simple interface is *for*, so it doesn't > > have to pull them from the formatted HTML of the /pypi interface. > >Just to be clear: the VCS URL is only ever used in the absence of any >other (suitable) package files? URLs are used when they match the criteria given by the user. If they put it on the command line explicitly, it gets used. If it's an obvious distribution filename at the end of the URL, the version info is used to determine whether it's relevant. Otherwise, it has to have a '#egg' tag to indicate what project and version the link is for... and then it'll get used if it matches the version spec that's being searched for. Usually, an SVN or other revision control link is going to have an '#egg=myproject-dev' tag, meaning "this is a URL for version 'dev' of 'myproject'", so it'll only be downloaded if it's the only link, OR if the user explicitly requests a version matching 'dev', and there's nothing with a higher version to match. > That's my reading of >http://peak.telecommunity.com/DevCenter/EasyInstall#downloading-and-installing-a-package The doc you want for that is this one, actually: http://peak.telecommunity.com/DevCenter/setuptools#making-your-package-available-for-easyinstall From martin at v.loewis.de Sun Apr 10 10:08:55 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Sun, 10 Apr 2011 10:08:55 +0200 Subject: [Catalog-sig] pypissh In-Reply-To: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> References: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> Message-ID: <4DA16597.8070205@v.loewis.de> > I tried pypissh 1.1. When I ran it with no arguments, it produced this > result: > > Traceback (most recent call last): > File "pypissh.py", line 85, in > f = urlopen('httpssh://submit at pypi.python.org/pypi') > File "pypissh.py", line 74, in urlopen > return _opener.open(req, data, timeout) > File "pypissh.py", line 57, in open > if req.get_full_url().startswith(_badprefix): > AttributeError: 'str' object has no attribute 'get_full_url' Thanks for the report. I hadn't tested __main__ in a while: I have now restored that in cd88f3c7fed9. There are apparently still interactions with the ssh output; unfortunately, I haven't been able to reproduce them after the first failure. Regards, Martin From stefan-usenet at bytereef.org Sun Apr 10 20:21:29 2011 From: stefan-usenet at bytereef.org (Stefan Krah) Date: Sun, 10 Apr 2011 20:21:29 +0200 Subject: [Catalog-sig] Download statistics for externally hosted packages Message-ID: <20110410182129.GA23072@sleipnir.bytereef.org> Hi, The download statistics for externally hosted packages are not accurate: http://pypi.python.org/pypi/cdecimal/2.2 This displays 0 downloads, while cdecimal has quite a healthy download rate, in particular from "high profile" users like banks, research institutes and stock exchanges. Would it be possible to replace the line with: "The main project site, which offers downloads, documentation and mailing lists, is hosted externally." I'm glad that the ratings are gone (even though cdecimal had a 5). I think it would be very nice to have the option of using PyPI as a simple catalog. By this I mean that I'd like to set up the package page once and not worry about new information (which might not be accurate) appearing on the page. Stefan Krah From martin at v.loewis.de Sun Apr 10 20:29:53 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Sun, 10 Apr 2011 20:29:53 +0200 Subject: [Catalog-sig] Download statistics for externally hosted packages In-Reply-To: <20110410182129.GA23072@sleipnir.bytereef.org> References: <20110410182129.GA23072@sleipnir.bytereef.org> Message-ID: <4DA1F721.10604@v.loewis.de> > The download statistics for externally hosted packages are not accurate: > > http://pypi.python.org/pypi/cdecimal/2.2 This was actually an accidental checkin - it was information from http://py3ksupport.appspot.com/pypi/cdecimal This was work in progress; I wanted to request from Brett an URL that only displays the Python 3 support information. In any case, I have reverted the checkin now. Regards, Martin From martin at v.loewis.de Sun Apr 10 20:47:35 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Sun, 10 Apr 2011 20:47:35 +0200 Subject: [Catalog-sig] pypissh In-Reply-To: <20110408181849.1992.529337496.divmod.xquotient.361@localhost.localdomain> References: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> <20110408181849.1992.529337496.divmod.xquotient.361@localhost.localdomain> Message-ID: <4DA1FB47.2070700@v.loewis.de> > Beyond this, if I try to use pypissh with a simple setup.py like this: > > import pypissh > pypissh.monkeypatch() > > from distutils.core import setup > > setup( > name="fake-experiment", version="1") > > It fails with this output: [...]> Submitting dist/fake-experiment-1.tar.gz to > http://submit at pypi.python.org/pypi > [Errno -2] Name or service not known I assume you have been using 2.5 or 2.6, right? I had tested it only with 2.7; in earlier versions, upload.py wasn't using urllib2. I have now added this support, and released 1.2, so please retry. Feedback is appreciated. Regards, Martin From stefan-usenet at bytereef.org Sun Apr 10 20:55:10 2011 From: stefan-usenet at bytereef.org (Stefan Krah) Date: Sun, 10 Apr 2011 20:55:10 +0200 Subject: [Catalog-sig] Download statistics for externally hosted packages In-Reply-To: <4DA1F721.10604@v.loewis.de> References: <20110410182129.GA23072@sleipnir.bytereef.org> <4DA1F721.10604@v.loewis.de> Message-ID: <20110410185510.GA23206@sleipnir.bytereef.org> "Martin v. L?wis" wrote: > > The download statistics for externally hosted packages are not accurate: > > > > http://pypi.python.org/pypi/cdecimal/2.2 > > This was actually an accidental checkin - it was information from > > http://py3ksupport.appspot.com/pypi/cdecimal > > This was work in progress; I wanted to request from Brett an URL > that only displays the Python 3 support information. In any case, > I have reverted the checkin now. Oh, sorry. - I was worried it might be a new feature that replaces the ratings. Thanks for all the work you are doing on PyPI! Stefan Krah From steve at holdenweb.com Mon Apr 11 14:57:24 2011 From: steve at holdenweb.com (Steve Holden) Date: Mon, 11 Apr 2011 08:57:24 -0400 Subject: [Catalog-sig] Please turn off ratings Message-ID: Ben Finney wrote: > "Martin v. L?wis" writes: > > > It seems to me that, given we're quite a long way into the potential > > user adoption of this feature, it's simply not being used. I would > > support removing it, as it seems that the only purpose it serves is > > to antagonise. > > > Thanks for your advise. I (now) believe this is a reasonable action > > to take, even though I really dislike not being able to defend what > > I believe is in the interest of the PyPI users. > > > Thank you for that decision, and for being honest about the difficulty > in making it. > And thank *you* for being, as far as I can tell, the only person on the requesting side of this issue appreciative enough to thank Martin for listening and responding. [obMonty] I thought this was Arguments, but it seems I ended up next door in Abuse ... This thread could probably be used as a model for how /not/ to conduct a debate on a technical issue over a public channel, and I would like all those involved to take it to heart as an object lesson in poor community collaboration. Sure, things like this will probably happen again, but let's just try and keep the heat down a little. I am aware that people (on both "sides") may come out of this smarting about the way (they perceive) others behaved. Let's just try and forget those feelings and agree to move on in some sort of comradely way if we can? On the technical side, it would seem logical to me to have third-party (community and commercial) sites establish ratings, and then have PyPI "scrape" the data (hopefully via established APIs) to present it in a friendly and helpful way to those users interacting through a browser. Thanks. PS: I do not frequent the list, so am asking Martin to forward this message if it doesn't make it through. -- Steve Holden steve at holdenweb.com From monitor at jacobian.org Wed Apr 13 03:36:37 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Tue, 12 Apr 2011 20:36:37 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection succeeded Message-ID: <1302658598.1@jacobian.org> Connection succeeded Service pypi.python.org Date: Tue, 12 Apr 2011 20:36:37 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[pypi.python.org:80] via TCP Your faithful employee, monit From exarkun at twistedmatrix.com Wed Apr 13 04:52:08 2011 From: exarkun at twistedmatrix.com (exarkun at twistedmatrix.com) Date: Wed, 13 Apr 2011 02:52:08 -0000 Subject: [Catalog-sig] Problem uploading an MSI to PyPI Message-ID: <20110413025208.1992.1121059104.divmod.xquotient.464@localhost.localdomain> Hello all, A member of the Tahoe-LAFS community was kind enough to build a Python 2.6 amd64 MSI installer for pyOpenSSL 0.12 for me (since I am not set up to build anything for 64 bit Windows). The MSI is available at http://twistedmatrix.com/~exarkun/pyOpenSSL-0.12 .win-amd64-py2.6.msi for the time being. It might be interesting to inspect in considering the question I have. First I tried uploading it to PyPI using the web interface. Unfortunately this resulted in a failed upload (several times) because "invalid distribution file". Then I tried using the distutils upload command (which I had to modify to allow uploading a file which was already built, rather than one which was being built at the moment). This produced the same result ("Upload failed (400): invalid distribution file"). So... is this a distutils bug? Is there something wrong with the MSI? Is it something wrong with pyOpenSSL's setup.py? Is it a bug in PyPI causing it to reject a valid MSI? (I hope the cross-posting is okay, since this topic seems to straddle these two areas.) Thanks, Jean-Paul From exarkun at twistedmatrix.com Wed Apr 13 05:00:17 2011 From: exarkun at twistedmatrix.com (exarkun at twistedmatrix.com) Date: Wed, 13 Apr 2011 03:00:17 -0000 Subject: [Catalog-sig] pypissh In-Reply-To: <4DA1FB47.2070700@v.loewis.de> References: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> <20110408181849.1992.529337496.divmod.xquotient.361@localhost.localdomain> <4DA1FB47.2070700@v.loewis.de> Message-ID: <20110413030017.1992.336972748.divmod.xquotient.466@localhost.localdomain> On 10 Apr, 06:47 pm, martin at v.loewis.de wrote: >>Beyond this, if I try to use pypissh with a simple setup.py like this: >> >> import pypissh >> pypissh.monkeypatch() >> >> from distutils.core import setup >> >> setup( >> name="fake-experiment", version="1") >> >>It fails with this output: >[...]> Submitting dist/fake-experiment-1.tar.gz to >>http://submit at pypi.python.org/pypi >>[Errno -2] Name or service not known > >I assume you have been using 2.5 or 2.6, right? I had tested it only >with 2.7; in earlier versions, upload.py wasn't using urllib2. > >I have now added this support, and released 1.2, so please retry. >Feedback is appreciated. Thanks. I was going to give this a try, but it looks like the wrong 1.2 release tarball is present on PyPI. Jean-Paul From martin at v.loewis.de Wed Apr 13 08:25:02 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 13 Apr 2011 08:25:02 +0200 Subject: [Catalog-sig] pypissh In-Reply-To: <20110413030017.1992.336972748.divmod.xquotient.466@localhost.localdomain> References: <20110408175837.1992.1187211369.divmod.xquotient.354@localhost.localdomain> <20110408181849.1992.529337496.divmod.xquotient.361@localhost.localdomain> <4DA1FB47.2070700@v.loewis.de> <20110413030017.1992.336972748.divmod.xquotient.466@localhost.localdomain> Message-ID: <4DA541BE.4070101@v.loewis.de> > Thanks. I was going to give this a try, but it looks like the wrong 1.2 > release tarball is present on PyPI. Sorry about this; sdist seems to have messed up. Please try again, with 1.3. Regards, Martin From monitor at jacobian.org Wed Apr 13 03:27:36 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Tue, 12 Apr 2011 20:27:36 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection failed Message-ID: <1302658059.1@jacobian.org> Connection failed Service pypi.python.org Date: Tue, 12 Apr 2011 20:27:35 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Thu Apr 14 22:23:48 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Thu, 14 Apr 2011 15:23:48 -0500 Subject: [Catalog-sig] [monit] d.pypi.python.org - Connection failed Message-ID: <1302812630.1@jacobian.org> Connection failed Service d.pypi.python.org Date: Thu, 14 Apr 2011 15:23:48 -0500 Action: alert Host: jacobian.org Description: failed, cannot open a connection to INET[d.pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Fri Apr 15 00:58:00 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Thu, 14 Apr 2011 17:58:00 -0500 Subject: [Catalog-sig] [monit] d.pypi.python.org - Connection succeeded Message-ID: <1302821881.1@jacobian.org> Connection succeeded Service d.pypi.python.org Date: Thu, 14 Apr 2011 17:58:00 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[d.pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Fri Apr 15 01:53:09 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Thu, 14 Apr 2011 18:53:09 -0500 Subject: [Catalog-sig] [monit] d.pypi.python.org - Connection failed Message-ID: <1302825190.1@jacobian.org> Connection failed Service d.pypi.python.org Date: Thu, 14 Apr 2011 18:53:09 -0500 Action: alert Host: jacobian.org Description: failed, cannot open a connection to INET[d.pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Fri Apr 15 08:53:27 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Fri, 15 Apr 2011 01:53:27 -0500 Subject: [Catalog-sig] [monit] d.pypi.python.org - Connection succeeded Message-ID: <1302850409.1@jacobian.org> Connection succeeded Service d.pypi.python.org Date: Fri, 15 Apr 2011 01:53:27 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[d.pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Fri Apr 15 09:27:57 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Fri, 15 Apr 2011 02:27:57 -0500 Subject: [Catalog-sig] [monit] d.pypi.python.org - Connection failed Message-ID: <1302852478.1@jacobian.org> Connection failed Service d.pypi.python.org Date: Fri, 15 Apr 2011 02:27:57 -0500 Action: alert Host: jacobian.org Description: failed, cannot open a connection to INET[d.pypi.python.org:80] via TCP Your faithful employee, monit From monitor at jacobian.org Fri Apr 15 11:30:42 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Fri, 15 Apr 2011 04:30:42 -0500 Subject: [Catalog-sig] [monit] d.pypi.python.org - Connection succeeded Message-ID: <1302859843.1@jacobian.org> Connection succeeded Service d.pypi.python.org Date: Fri, 15 Apr 2011 04:30:42 -0500 Action: alert Host: jacobian.org Description: connection succeeded to INET[d.pypi.python.org:80] via TCP Your faithful employee, monit From techtonik at gmail.com Sat Apr 16 16:10:27 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Sat, 16 Apr 2011 17:10:27 +0300 Subject: [Catalog-sig] PyPI Message-ID: Hi Richard, Sorry that I can't reply to your r899 commit directly, but did you forget to link About page on the site? I can't find it is being referenced anywhere. Please, CC. -- anatoly t. From techtonik at gmail.com Sat Apr 16 16:20:35 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Sat, 16 Apr 2011 17:20:35 +0300 Subject: [Catalog-sig] Remove rating vs turn off Message-ID: I see that ratings were removed from source code base, not turned off. Why? I like the idea to have my own _private_ ratings, flags, keyword and list of important bugs for package versions and sharing it with other people via OpenSocial API, for example. Please, CC. -- anatoly t. From techtonik at gmail.com Sat Apr 16 16:42:29 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Sat, 16 Apr 2011 17:42:29 +0300 Subject: [Catalog-sig] [PATCH] Prettify JSON output Message-ID: Hi, Right now JSON output is not really suitable for development. This patch improves the outlook. It also adds inline content-disposition header as an attempt to fix FF4 behavior which still downloads the file unlike Chromium. Please, CC. -- anatoly t. -------------- next part -------------- A non-text attachment was scrubbed... Name: pretty.json.pypi.patch Type: application/octet-stream Size: 1153 bytes Desc: not available URL: From tjreedy at udel.edu Sat Apr 16 19:28:44 2011 From: tjreedy at udel.edu (Terry Reedy) Date: Sat, 16 Apr 2011 13:28:44 -0400 Subject: [Catalog-sig] [PATCH] Prettify JSON output In-Reply-To: References: Message-ID: <4DA9D1CC.7010200@udel.edu> On 4/16/2011 10:42 AM, anatoly techtonik wrote: > Please, CC. Anatoly, you can follow this list as newsgroup gmane.comp.python.catalog at gmane.news.org. with any decent mail/news program, like Thunderbird. Just subscribe and unsubscribe as you wish. -- Terry Jan Reedy From techtonik at gmail.com Sat Apr 16 20:45:10 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Sat, 16 Apr 2011 21:45:10 +0300 Subject: [Catalog-sig] [PATCH] Prettify JSON output In-Reply-To: <4DA9D1CC.7010200@udel.edu> References: <4DA9D1CC.7010200@udel.edu> Message-ID: Thanks for the hint, but I prefer web tools to installable software, because I use several computers. Saves me a lot of time. I can tune GMail, but I already have a lot of filters that stand in my way. All these lists are cluttering my archive search. So it would be nice if I could just add Google Groups frontend for Mailman. Less distractions, selective subscribe, saved read/unread status and, of course search that works regardless of my subscription status and PC. http://mail.python.org/pipermail/mailman-users/2011-March/071335.html -- anatoly t. On Sat, Apr 16, 2011 at 8:28 PM, Terry Reedy wrote: > On 4/16/2011 10:42 AM, anatoly techtonik wrote: > >> Please, CC. > > Anatoly, you can follow this list as newsgroup gmane.comp.python.catalog at > gmane.news.org. > with any decent mail/news program, like Thunderbird. > Just subscribe and unsubscribe as you wish. > > -- > Terry Jan Reedy > From pydanny at gmail.com Mon Apr 18 19:19:03 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Mon, 18 Apr 2011 10:19:03 -0700 Subject: [Catalog-sig] Some minor issues with page content Message-ID: Hey guys, I get a 404 when I go here which is off the package upload page: http://www.python.org/doc/dist/package-upload.html Also, when I go to the tutorial I see instructions pointing at easy_install. I thought the move was towards Pip. See http://wiki.python.org/moin/CheeseShopTutorial Thanks so much for all the hard work! -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From pydanny at gmail.com Mon Apr 18 19:29:01 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Mon, 18 Apr 2011 10:29:01 -0700 Subject: [Catalog-sig] How about a dedicated web service mirror? Message-ID: Here is something that we all might find useful. When Python Packages launches it will be hitting the XMLRPC server over 10,000 times a day - at least once per listed package. Django Packages already hits PyPI about 3000 times a day. We'll be doing this in a distributed task queu but the fact remains that the server is going to be hit a decent amount. Because of the relative ease of setting up the Packaginator tool these sites are building other people will also be using this same API. Or maybe they'll build their own system and do even more. So maybe dedicated mirrors that truncate the web UI and just provide the web service API? Cheers, -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From mal at egenix.com Mon Apr 18 20:20:54 2011 From: mal at egenix.com (M.-A. Lemburg) Date: Mon, 18 Apr 2011 20:20:54 +0200 Subject: [Catalog-sig] Some minor issues with page content In-Reply-To: References: Message-ID: <4DAC8106.5080000@egenix.com> Daniel Greenfeld wrote: > Hey guys, > > I get a 404 when I go here which is off the package upload page: > http://www.python.org/doc/dist/package-upload.html > > Also, when I go to the tutorial I see instructions pointing at > easy_install. I thought the move was towards Pip. See > http://wiki.python.org/moin/CheeseShopTutorial Feel free to add a new section on pip. AFAIK, pip only supports source-code packages (please correct if this is no longer the case), so easy_install is still the way to go for binary packages. -- Marc-Andre Lemburg eGenix.com Professional Python Services directly from the Source (#1, Apr 18 2011) >>> 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 http://www.egenix.com/company/contact/ From mal at egenix.com Mon Apr 18 20:29:54 2011 From: mal at egenix.com (M.-A. Lemburg) Date: Mon, 18 Apr 2011 20:29:54 +0200 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: References: Message-ID: <4DAC8322.2000407@egenix.com> Daniel Greenfeld wrote: > Here is something that we all might find useful. > > When Python Packages launches it will be hitting the XMLRPC server > over 10,000 times a day - at least once per listed package. Django > Packages already hits PyPI about 3000 times a day. We'll be doing this > in a distributed task queu but the fact remains that the server is > going to be hit a decent amount. Is that really necessary ? PyPI has a mechanism to subscribe to recent changes only, which should be enough for the needs of the sites you mention. > Because of the relative ease of setting up the Packaginator tool these > sites are building other people will also be using this same API. Or > maybe they'll build their own system and do even more. > > So maybe dedicated mirrors that truncate the web UI and just provide > the web service API? -- Marc-Andre Lemburg eGenix.com Professional Python Services directly from the Source (#1, Apr 18 2011) >>> 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 http://www.egenix.com/company/contact/ From martin at v.loewis.de Mon Apr 18 21:50:46 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Mon, 18 Apr 2011 21:50:46 +0200 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: References: Message-ID: <4DAC9616.3020200@v.loewis.de> > Here is something that we all might find useful. > > When Python Packages launches it will be hitting the XMLRPC server > over 10,000 times a day - at least once per listed package. Django > Packages already hits PyPI about 3000 times a day. We'll be doing this > in a distributed task queu but the fact remains that the server is > going to be hit a decent amount. I'd like to second Marc-Andre's question: why is it that you will need to make 10,000 queries a day? Regards, Martin From pydanny at gmail.com Tue Apr 19 07:22:13 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Mon, 18 Apr 2011 22:22:13 -0700 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: <4DAC9F5E.4060109@v.loewis.de> References: <4DAC9616.3020200@v.loewis.de> <4DAC9F5E.4060109@v.loewis.de> Message-ID: Martin, Thanks for this link but I'm not sure about what the entries mean. Using some entries from SqlAlchemy as an example: SQLAlchemy,SQLAlchemy-0.6beta2.tar.gz,z3c.pypimirror/1.0.15.1,3 SQLAlchemy,SQLAlchemy-0.6beta3.tar.gz,setuptools/0.6c11,1 SQLAlchemy,SQLAlchemy-0.6beta3.tar.gz,setuptools/0.6c9,1 SQLAlchemy,SQLAlchemy-0.6beta3.tar.gz,z3c.pypimirror/1.0.15.1,3 I guess the first column is the package identifier, the second is the file + ???, and the fourth column the download count? Am I close? Is there any formal documentation for this portion of PyPI? I'm know this is easier on PyP's server but I'm hesitant to interact with an undocumented data source. If documentation doesn't exist for it, I'm happy enough to write the formal documentation once I understand it. Also, other tools are using the XMLRPC API to make large requests against PyPI. http://pypi.python.org/pypi/vanity and http://pypi.appspot.com come to mind. Thanks! Danny On Mon, Apr 18, 2011 at 1:30 PM, "Martin v. L?wis" wrote: > Am 18.04.2011 22:20, schrieb Daniel Greenfeld: >> All the packages need to have their download numbers checked for metrics. > > Please do that incrementally, downloading a single file per day from > > http://pypi.python.org/stats/days/ > > All files displayed there have been incorporated in the total stats, > so if you download the total stats once, then you can start downloading > the incremental changes the next day. Notice that some of the mirrors > are a few days behind in generating these files, so even the older files > may still change after first being generated (so you may need to fetch > three or so files, or the monthly one). > > Regards, > Martin > -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From martin at v.loewis.de Tue Apr 19 08:09:48 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 19 Apr 2011 08:09:48 +0200 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: References: <4DAC9616.3020200@v.loewis.de> <4DAC9F5E.4060109@v.loewis.de> Message-ID: <4DAD272C.10800@v.loewis.de> > Thanks for this link but I'm not sure about what the entries mean. > Using some entries from SqlAlchemy as an example: > > SQLAlchemy,SQLAlchemy-0.6beta2.tar.gz,z3c.pypimirror/1.0.15.1,3 > SQLAlchemy,SQLAlchemy-0.6beta3.tar.gz,setuptools/0.6c11,1 > SQLAlchemy,SQLAlchemy-0.6beta3.tar.gz,setuptools/0.6c9,1 > SQLAlchemy,SQLAlchemy-0.6beta3.tar.gz,z3c.pypimirror/1.0.15.1,3 > > I guess the first column is the package identifier, the second is the > file + ???, and the fourth column the download count? Am I close? All correct; the third one is the user-agent that performed the download. > Is there any formal documentation for this portion of PyPI? Yes, see http://www.python.org/dev/peps/pep-0381/#statistics-page > I'm know > this is easier on PyP's server but I'm hesitant to interact with an > undocumented data source. If documentation doesn't exist for it, I'm > happy enough to write the formal documentation once I understand it. It's certainly an official interface, and this dataset is actually also intended for applications like yours (the local-stats pages are only intended for the mirroring infrastructure itself). > Also, other tools are using the XMLRPC API to make large requests > against PyPI. http://pypi.python.org/pypi/vanity and > http://pypi.appspot.com come to mind. Such applications are encouraged to use batched XML-RPC. As long as they aren't causing load problems on the server, I'm fine with the tools doing what they do - it's just that batching would actually get them the results faster. However, pypi.appspot.com does *not* make heavy requests through XML-RPC. It is b.pypi.python.org, and most certainly uses the journal to only request recent changes. Regards, Martin From alexis at notmyidea.org Tue Apr 19 09:45:57 2011 From: alexis at notmyidea.org (=?windows-1252?Q?Alexis_M=E9taireau?=) Date: Tue, 19 Apr 2011 09:45:57 +0200 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: <4DAD272C.10800@v.loewis.de> References: <4DAC9616.3020200@v.loewis.de> <4DAC9F5E.4060109@v.loewis.de> <4DAD272C.10800@v.loewis.de> Message-ID: <4DAD3DB5.4010500@notmyidea.org> On 19/04/2011 08:09, "Martin v. L?wis" wrote: > It's certainly an official interface, and this dataset is actually > also intended for applications like yours (the local-stats pages > are only intended for the mirroring infrastructure itself). Is there some statistics about the XML-RPC interface available somewhere (so we can put in perspective the use of XML-RPC and the use of simple index)? In packaging we prefered the use of the simple API over the XML-RPC interface but in some edge-cases we need to rely on XML-RPC (like search on something else than the name of the project). Because XML-RPC is not mirrored as the simple API is, this can sometimes block the install process. Having a way to mirror it can be a way to avoid such blocking behaviour when in case the central PyPI server goes down. -- Alexis ? http://notmyidea.org From merwok at netwok.org Tue Apr 19 18:11:49 2011 From: merwok at netwok.org (=?UTF-8?Q?=C3=89ric_Araujo?=) Date: Tue, 19 Apr 2011 18:11:49 +0200 Subject: [Catalog-sig] =?utf-8?q?How_about_a_dedicated_web_service_mirror?= =?utf-8?q?=3F?= In-Reply-To: References: Message-ID: Hi, > When Python Packages launches it will be hitting the XMLRPC server > over 10,000 times a day - at least once per listed package. BTW, what is Python Packages? I?ve looked at Django Packages and it looks like a project index, like PyPI, so I wonder what Python Packages is that PyPI isn?t (or couldn?t be patched to be). (The term ?Python Packages? in a search engine of course matches documentation for Python packages (things you can import, not things you distribute), so I?m asking here.) Regards From martin at v.loewis.de Tue Apr 19 20:32:05 2011 From: martin at v.loewis.de (=?windows-1252?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Tue, 19 Apr 2011 20:32:05 +0200 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: <4DAD3DB5.4010500@notmyidea.org> References: <4DAC9616.3020200@v.loewis.de> <4DAC9F5E.4060109@v.loewis.de> <4DAD272C.10800@v.loewis.de> <4DAD3DB5.4010500@notmyidea.org> Message-ID: <4DADD525.5030704@v.loewis.de> > Is there some statistics about the XML-RPC interface available somewhere > (so we can put in perspective the use of XML-RPC and the use of simple > index)? Not directly. However, take a look at http://pypi.python.org/webstats/usage_201104.html Accesses to /pypi include XML-RPC calls. In addition, looking at the number of accesses by user-agent give a clue on how much programmatic accesses are made. Regards, Martin From jacob at jacobian.org Tue Apr 19 22:36:36 2011 From: jacob at jacobian.org (Jacob Kaplan-Moss) Date: Tue, 19 Apr 2011 15:36:36 -0500 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: References: Message-ID: On Tue, Apr 19, 2011 at 11:11 AM, ?ric Araujo wrote: > BTW, what is Python Packages? ?I?ve looked at Django Packages and it > looks like a project index, like PyPI, so I wonder what Python Packages > is that PyPI isn?t (or couldn?t be patched to be). I can't speak for Danny's motivation behind Python/Django Packages, but I can give you some insights into why Django Packages has been valuable for our community. Hopefully that can answer your questions a bit: We've been really lucky to have a awesome community of third-party Django apps, but today we've got a classic paradox of choice problem. PyPI lists about 1,100 Django packages! This is clearly a good thing, but it's also a problem: it's hard to figure out what to use for any particular need. PyPI's got a bit of an interface for browsing these packages (http://pypi.python.org/pypi?:action=browse&c=523), and there's a search engine, but you still end up with a big choice problem. If I search for "django media" on PyPI, I find dozens of packages that can handle my media-file needs... but should I use django-mediafiles or django-staticmedia or django-fresh-media or... The key insight that (for me) makes Django Packages useful is that it *groups* related packages together and lets me see, at a glance, which of a set of related packages is the most useful. So I can visit http://djangopackages.com/grids/g/asset-managers/ and discover that django-staticfiles, while the most popular (by an "I use this" count) doesn't support, say, JavaScript compression. If that's a feature I need, then I can use that package. I've started pointing people to grids (groups of packages) on Django Packages because they're living documents that describe all the solutions in a given problem space. I don't have to worry about my advice getting out of date, and I don't have to force my choices onto people. I can simply point them to the comparison pages and let them draw their own conclusions. Again speaking strictly as a user, I would *love* this to exist for the greater Python ecosystem. I would really like to point someone to http://pythonpackages.com/grids/g/xml-processors/ when helping them decide between an XML engine. Or a web framework, for that matter! Perhaps this could be part of some future PyPI. However, I see two reasons why it probably won't: * Technically, the fact that PyPI's written as a hand-rolled WSGI app limits the number of people who're going to want to contribute. Packaginator (the code behind * Packages) lists 32 contributors already. Would those people have learned the PyPI codebase? I doubt it. I certainly am a hell of a lot more likely to contribute to a project written in Django, or Flask, or Pylons, or Pryamid than one where I have to take time re-learn a bunch of boring stuff like templates and db abstractions. With apologies to Aaron Sorkin, If Packaginator could have been part of PyPI, it would be part of PyPI. * Socially, Python Packages, like all community-generated content, will always be a bit of a flawed product. Since the data there is mostly human-driven, it's never going to be comprehensive, complete, or accurate. Again taking the Django asset manager grid as an example, I can see at least two things that are missing there. This doesn't bother me as a user of Django Packages -- just as the parallel issue doesn't bother me as a user of Wikipedia. But I would *not* want to make Django Packages part of Django's official web presence because that would imply an authority of the results that could never actually be matched. As a community project, Python Packages can avoid the political problems making it part of PyPI would entail. No endorsements are needed or implied. To sum up, I'm incredibly happy with Django Packages, and that's why I'm waiting anxiously for Python Packages to launch (and why I'm helping make it happen). Jacob From richard at python.org Wed Apr 20 02:03:02 2011 From: richard at python.org (Richard Jones) Date: Wed, 20 Apr 2011 10:03:02 +1000 Subject: [Catalog-sig] How about a dedicated web service mirror? In-Reply-To: References: Message-ID: On Wed, Apr 20, 2011 at 6:36 AM, Jacob Kaplan-Moss wrote: > * Technically, the fact that PyPI's written as a hand-rolled WSGI app > limits the number of people who're going to want to contribute. > Packaginator (the code behind * Packages) lists 32 contributors > already. Would those people have learned the PyPI codebase? I doubt > it. I certainly am a hell of a lot more likely to contribute to a > project written in Django, or Flask, or Pylons, or Pryamid than one > where I have to take time re-learn a bunch of boring stuff like > templates and db abstractions. > > With apologies to Aaron Sorkin, If Packaginator could have been part > of PyPI, it would be part of PyPI. As the guy who wrote that hand-rolled WSGI app in the first place (back when Zope was really the only thing around) I've said a number of times that I'd love for the code to be modernised. I was even looking forward to making a start on it at the US PyCon sprints this year, but alas my employer couldn't fund my attendance. Packaginator sounds like it could have been a viable replacement. But they're going their own way, like so many other projects that have anything to do with PyPI. > * Socially, Python Packages, like all community-generated content, > will always be a bit of a flawed product. Since the data there is > mostly human-driven, it's never going to be comprehensive, complete, > or accurate. Again taking the Django asset manager grid as an example, > I can see at least two things that are missing there. > > This doesn't bother me as a user of Django Packages -- just as the > parallel issue doesn't bother me as a user of Wikipedia. But I would > *not* want to make Django Packages part of Django's official web > presence because that would imply an authority of the results that > could never actually be matched. As a community project, Python > Packages can avoid the political problems making it part of PyPI would > entail. No endorsements are needed or implied. And let's not downplay this. PyPI will gain no fancy new user-generated content features because Martin and I both know that someone, somewhere will complain and we'll get a load of crap for it. A separate project on a separate website has no such fears. So we'll just focus on the core stuff and hope that we can provide useful services to those websites. Here's hoping that Packaginator really takes off and we can get some really cool extra data around PyPI packages to help end developers out. And maybe there could be some kernel in there that we might adapt to become the new PyPI implementation... Richard From richard at python.org Wed Apr 20 02:11:25 2011 From: richard at python.org (Richard Jones) Date: Wed, 20 Apr 2011 10:11:25 +1000 Subject: [Catalog-sig] [PATCH] Prettify JSON output In-Reply-To: References: Message-ID: On Sun, Apr 17, 2011 at 12:42 AM, anatoly techtonik wrote: > Right now JSON output is not really suitable for development. This > patch improves the outlook. It also adds inline content-disposition > header as an attempt to fix FF4 behavior which still downloads the > file unlike Chromium. This patch has now been applied. Richard From richard at python.org Wed Apr 20 02:12:21 2011 From: richard at python.org (Richard Jones) Date: Wed, 20 Apr 2011 10:12:21 +1000 Subject: [Catalog-sig] PyPI In-Reply-To: References: Message-ID: On Sun, Apr 17, 2011 at 12:10 AM, anatoly techtonik wrote: > Sorry that I can't reply to your r899 commit directly, but did you > forget to link About page on the site? I can't find it is being > referenced anywhere. The page is not complete yet. From richard at python.org Wed Apr 20 02:14:30 2011 From: richard at python.org (Richard Jones) Date: Wed, 20 Apr 2011 10:14:30 +1000 Subject: [Catalog-sig] Remove rating vs turn off In-Reply-To: References: Message-ID: On Sun, Apr 17, 2011 at 12:20 AM, anatoly techtonik wrote: > I see that ratings were removed from source code base, not turned off. > Why? I like the idea to have my own _private_ ratings, flags, keyword > and list of important bugs for package versions and sharing it with > other people via OpenSocial API, for example. Please submit this kind of idea to the Packaginator or similar projects. Richard From techtonik at gmail.com Wed Apr 20 09:48:26 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Wed, 20 Apr 2011 10:48:26 +0300 Subject: [Catalog-sig] [PATCH] Prettify JSON output In-Reply-To: References: Message-ID: On Wed, Apr 20, 2011 at 3:11 AM, Richard Jones wrote: > On Sun, Apr 17, 2011 at 12:42 AM, anatoly techtonik wrote: >> Right now JSON output is not really suitable for development. This >> patch improves the outlook. It also adds inline content-disposition >> header as an attempt to fix FF4 behavior which still downloads the >> file unlike Chromium. > > This patch has now been applied. Thanks. It doesn't fix FF, but at least there is now a bug report for it (besides that JSONView is a very nice dev add on to be installed). https://bugzilla.mozilla.org/show_bug.cgi?id=651439 By the way, there is an extra change slipped in this commit. -- anatoly t. From pydanny at gmail.com Wed Apr 20 15:44:29 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Wed, 20 Apr 2011 06:44:29 -0700 Subject: [Catalog-sig] Catalog-SIG Digest, Vol 86, Issue 33 In-Reply-To: References: Message-ID: > Message: 1 > Date: Wed, 20 Apr 2011 10:14:30 +1000 > From: Richard Jones > To: anatoly techtonik > Cc: catalog-sig > Subject: Re: [Catalog-sig] Remove rating vs turn off > Message-ID: > Content-Type: text/plain; charset=ISO-8859-1 > > On Sun, Apr 17, 2011 at 12:20 AM, anatoly techtonik wrote: >> I see that ratings were removed from source code base, not turned off. >> Why? I like the idea to have my own _private_ ratings, flags, keyword >> and list of important bugs for package versions and sharing it with >> other people via OpenSocial API, for example. > > Please submit this kind of idea to the Packaginator or similar projects. > > > ? ? Richard Anatoly, Richard is right. Packaginator lets you stand up your own version of the tool, which merely grabs from PyPI various stats. While Packaginator core will never include a private rating system (everything is based on hard metrics because ratings are too easily skewed), we are working on an API so you can add that sort of functionality to your own instance without forking so hard you can't accept patches. However, as for integrating the complexity of OpenSocial API, thats something you can do in your own fork. -- 'Knowledge is Power' Daniel Greenfeld http://pydanny.com http://cartwheelweb.com From pydanny at gmail.com Wed Apr 20 16:15:59 2011 From: pydanny at gmail.com (Daniel Greenfeld) Date: Wed, 20 Apr 2011 07:15:59 -0700 Subject: [Catalog-sig] Catalog-SIG Digest, Vol 86, Issue 32 In-Reply-To: References: Message-ID: Date: Wed, 20 Apr 2011 10:03:02 +1000 From: Richard Jones To: Jacob Kaplan-Moss Cc: catalog-sig Subject: Re: [Catalog-sig] How about a dedicated web service mirror? Message-ID: Content-Type: text/plain; charset=ISO-8859-1 >As the guy who wrote that hand-rolled WSGI app in the first place >(back when Zope was really the only thing around) I've said a number >of times that I'd love for the code to be modernised. I was even l>ooking forward to making a start on it at the US PyCon sprints this >year, but alas my employer couldn't fund my attendance. Packaginator >sounds like it could have been a viable replacement. But they're going >their own way, like so many other projects that have anything to do >with PyPI. The three big reasons why we never added PyPI like capability to Django/Python Packages (now Packaginator) were: 1. A lot of people we respect said it would be best to support PyPI by providing accessory functionality rather than replacing it. 2. Adding package handling components to the project would take more time then we had to contribute to the project. 3. At PyCon people from various other groups (Plone, Pyramid, Fedora, Ubuntu, Vim) wanted to use it for their own needs. 4. We never wanted to be in the hotseat for when our package system went down. >> * Socially, Python Packages, like all community-generated content, >> will always be a bit of a flawed product. Since the data there is >> mostly human-driven, it's never going to be comprehensive, complete, >> or accurate. Again taking the Django asset manager grid as an example, >> I can see at least two things that are missing there. Jacob, why haven't you added them? :-) >> This doesn't bother me as a user of Django Packages -- just as the >> parallel issue doesn't bother me as a user of Wikipedia. But I would >> *not* want to make Django Packages part of Django's official web >> presence because that would imply an authority of the results that >> could never actually be matched. As a community project, Python >> Packages can avoid the political problems making it part of PyPI would >> entail. No endorsements are needed or implied. Yup. We are happy to give up the data to the DSF wants at any time but I'm so very glad we never pursued DSF funding and that they've been hands-off. > And let's not downplay this. PyPI will gain no fancy new > user-generated content features because Martin and I both know that > someone, somewhere will complain and we'll get a load of crap for it. We are actually worried about Python Packages in this regard. Which is why we locked down the grid permissions system for when we launch that project. The grids and packages will be moderated by formal moderators, but Package owners can edit content related to their own work. Also, all the Packaginator metadata fields are based off of hard metrics. There is no rating system on the project. The only thing that shows up in metadata displays that is directly user-entered is a single boolean of whether or not you use a particular package. > A separate project on a separate website has no such fears. So we'll > just focus on the core stuff and hope that we can provide useful > services to those websites. We really appreciate the hard work you've put into PyPI. PyPI, being what it is, provides an absolutely critical service to the Python community. Hosting packages is hard work and should be the ultimate focus of the project. Everything else (docs, evaluations, commentary) distracts from that focus. > Here's hoping that Packaginator really takes off and we can get some > really cool extra data around PyPI packages to help end developers > out. And maybe there could be some kernel in there that we might adapt > to become the new PyPI implementation... Indeed. Lets see how things play out. Daniel Greenfeld From techtonik at gmail.com Tue Apr 26 19:02:23 2011 From: techtonik at gmail.com (anatoly techtonik) Date: Tue, 26 Apr 2011 20:02:23 +0300 Subject: [Catalog-sig] reST not detected Message-ID: Hi, I've uploaded my module, but documentation is not rendered as restructured text. Why? http://pypi.python.org/pypi/pager/0.1 Please, CC. -- anatoly t. From lists at zopyx.com Tue Apr 26 19:33:43 2011 From: lists at zopyx.com (Andreas Jung) Date: Tue, 26 Apr 2011 19:33:43 +0200 Subject: [Catalog-sig] reST not detected In-Reply-To: References: Message-ID: <4DB701F7.4010009@zopyx.com> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Please check your reST yourself: suxmac2:pager-0.1 ajung$ rst2html.py README.rst >out README.rst:46: (ERROR/3) Unknown interpreted text role "func". - -aj anatoly techtonik wrote: > Hi, > > I've uploaded my module, but documentation is not rendered as > restructured text. Why? > http://pypi.python.org/pypi/pager/0.1 > > Please, CC. > -- > anatoly t. > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig - -- ZOPYX Limited | zopyx group Charlottenstr. 37/1 | The full-service network for Zope & Plone D-72070 T?bingen | Produce & Publish www.zopyx.com | www.produce-and-publish.com - ------------------------------------------------------------------------ E-Publishing, Python, Zope & Plone development, Consulting -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQGUBAEBAgAGBQJNtwH2AAoJEADcfz7u4AZj0eoLv3DCMGkoocfjVAWDd6Rr0VHT GXcPIa7LK1B3ID43J93xK94vZ7hsbYqrpDPEI6sctqg/AMTOO90M+VTg/CKENSnf mJkgQN/Ti9Uf9gyKycMBSD4gXFapqi7m1RXetFL2sjCQhWftuFVT6ZDpmeQxaLlm xDWM4GN3vadlo/cpWRJIqaSfRFrxiVrFt2XBvmR/8PIwPDBaXcycJPFIsDqCrXdc QV0Eok/ERn2h3Ll3ojZ9dJX9p8xvwFVwaOb+jPXJQI9chAytrK5lLBwUIV+OfbtX ggEYHnx7phcXa138yM1RqIqGf4YQq5ANLsgUXm/pys/ZOgmnfbO9tyMG5mYo6D4P GRKt+Rf2EkMvgPvvsy+wJALKyTC62dlp9rbGtPPNXrR8VQNA8A49Y7qA2WFxTVHu 1OlGUFt/O65f8hGnZEna7BRfc7JwnzZHk+NW9AoomsiGbwAbtT6FnRCG/Lk2IlKv XIxbNQ1a0rK4U2NxKKzZQp8nux9d+5s= =D6YH -----END PGP SIGNATURE----- -------------- next part -------------- A non-text attachment was scrubbed... Name: lists.vcf Type: text/x-vcard Size: 316 bytes Desc: not available URL: From benji at benjiyork.com Tue Apr 26 19:35:54 2011 From: benji at benjiyork.com (Benji York) Date: Tue, 26 Apr 2011 13:35:54 -0400 Subject: [Catalog-sig] reST not detected In-Reply-To: References: Message-ID: On Tue, Apr 26, 2011 at 1:02 PM, anatoly techtonik wrote: > I've uploaded my module, but documentation is not rendered as > restructured text. Why? > http://pypi.python.org/pypi/pager/0.1 It was probably the "func" role in "Default :func:`prompt` callback". % rst2html /tmp/1 > /dev/null /tmp/1:45: (ERROR/3) Unknown interpreted text role "func". -- Benji York From monitor at jacobian.org Wed Apr 27 05:49:40 2011 From: monitor at jacobian.org (monitor at jacobian.org) Date: Tue, 26 Apr 2011 22:49:40 -0500 Subject: [Catalog-sig] [monit] pypi.python.org - Connection failed Message-ID: <1303876184.1@jacobian.org> Connection failed Service pypi.python.org Date: Tue, 26 Apr 2011 22:49:40 -0500 Action: alert Host: jacobian.org Description: failed protocol test [HTTP] at INET[pypi.python.org:80] via TCP Your faithful employee, monit From martin at v.loewis.de Wed Apr 27 20:39:32 2011 From: martin at v.loewis.de (=?ISO-8859-15?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Wed, 27 Apr 2011 20:39:32 +0200 Subject: [Catalog-sig] PyPI mirror key rollover Message-ID: <4DB862E4.9060207@v.loewis.de> I came up with a key rollover scheme for the server key on PyPI. The objective of this key rollover is to protect against brute-force attacks of people trying to crack the key. If the main server itself gets compromised (and the private key leaks), this scheme will not help, and we will need to reset the package database back to a known-safe point in time. The idea is this: At any point in time, two keys are known - the "current" key, and the "new" key, https://pypi.python.org/serverkey https://pypi.python.org/serverkey.new There will be a key rollover every year on July 1st. At this point, the "new" key becomes current, and a new "new" key is generated and published. Tools using the mirror infrastructure are encouraged to hard-code (cache) both keys in their code base, and attempt validation of a "simple" page against each of them. If both validations fail, either the keys are outdated, or the mirror they are using has been compromised. In this case, they should fall back to not using mirrors, but contact the master directly. After a key rollover, mirrors will need some time to refetch the signatures for all "simple" pages, so the old current key may be needed some time after the key rollover still. Within one day, all mirrors should have caught up. Also after a key rollover, tools should update their cached copies of the keys, and produce new releases. Users will then have one year to upgrade their tools to continue to use the mirror infrastructure. Tools may also chose to offer users to update the serverkeys (after due validation), or may chose to download them automatically, in which case they should validate the https server certicate for authenticity of the keys. The key rollover will be logged in the PyPI journal, using an empty package name and an empty release. TOOLS USING THE JOURNAL MAY NEED TO BE FIXED TO ACCOMMODATE EMPTY PACKAGE NAMES. Earlier today, such a journal entry was already added; I took it out again when I noticed that some tools actually do need to be fixed. Unless some flaws are detected in this scheme, I'll do the first key rollover this coming July. Regards, Martin From mal at egenix.com Thu Apr 28 10:26:41 2011 From: mal at egenix.com (M.-A. Lemburg) Date: Thu, 28 Apr 2011 10:26:41 +0200 Subject: [Catalog-sig] PyPI mirror key rollover In-Reply-To: <4DB862E4.9060207@v.loewis.de> References: <4DB862E4.9060207@v.loewis.de> Message-ID: <4DB924C1.600@egenix.com> "Martin v. L?wis" wrote: > I came up with a key rollover scheme for the server key on PyPI. > [...] > > The key rollover will be logged in the PyPI journal, > using an empty package name and an empty release. TOOLS USING > THE JOURNAL MAY NEED TO BE FIXED TO ACCOMMODATE EMPTY PACKAGE > NAMES. Earlier today, such a journal entry was already added; > I took it out again when I noticed that some tools actually > do need to be fixed. I can't comment on the other parts of the proposal, but the above suggestions doesn't sound like a good solution: an empty package name in the update stream looks more like a server or client decoding bug than a trigger to do a key update. Wouldn't it be better to use a descriptive package name such as "pypi-serverkey-update" together with a package version which identifies the new serverkey version as trigger ? -- Marc-Andre Lemburg eGenix.com Professional Python Services directly from the Source (#1, Apr 28 2011) >>> Python/Zope Consulting and Support ... http://www.egenix.com/ >>> mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ >>> mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/ ________________________________________________________________________ 2011-06-20: EuroPython 2011, Florence, Italy 53 days to go ::: 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 http://www.egenix.com/company/contact/ From jannis at leidel.info Thu Apr 28 11:06:53 2011 From: jannis at leidel.info (Jannis Leidel) Date: Thu, 28 Apr 2011 11:06:53 +0200 Subject: [Catalog-sig] PyPI mirror key rollover In-Reply-To: <4DB924C1.600@egenix.com> References: <4DB862E4.9060207@v.loewis.de> <4DB924C1.600@egenix.com> Message-ID: <5B5AB171-A564-49DA-AF86-CEB78103CFDA@leidel.info> On 28.04.2011, at 10:26, M.-A. Lemburg wrote: > "Martin v. L?wis" wrote: >> I came up with a key rollover scheme for the server key on PyPI. >> [...] >> >> The key rollover will be logged in the PyPI journal, >> using an empty package name and an empty release. TOOLS USING >> THE JOURNAL MAY NEED TO BE FIXED TO ACCOMMODATE EMPTY PACKAGE >> NAMES. Earlier today, such a journal entry was already added; >> I took it out again when I noticed that some tools actually >> do need to be fixed. > > I can't comment on the other parts of the proposal, but the above > suggestions doesn't sound like a good solution: an empty package > name in the update stream looks more like a server or client > decoding bug than a trigger to do a key update. > > Wouldn't it be better to use a descriptive package name such > as "pypi-serverkey-update" together with a package version > which identifies the new serverkey version as trigger ? +1 Yeah, a convention like that seems better than an empty release. Jannis From ziade.tarek at gmail.com Thu Apr 28 11:27:14 2011 From: ziade.tarek at gmail.com (=?ISO-8859-1?Q?Tarek_Ziad=E9?=) Date: Thu, 28 Apr 2011 11:27:14 +0200 Subject: [Catalog-sig] PyPI mirror key rollover In-Reply-To: <4DB862E4.9060207@v.loewis.de> References: <4DB862E4.9060207@v.loewis.de> Message-ID: On Wed, Apr 27, 2011 at 8:39 PM, "Martin v. L?wis" wrote: > I came up with a key rollover scheme for the server key on PyPI. > The objective of this key rollover is to protect against brute-force > attacks of people trying to crack the key. If the main server itself > gets compromised (and the private key leaks), this scheme will not > help, and we will need to reset the package database back to a > known-safe point in time. > > The idea is this: At any point in time, two keys are known - the > "current" key, and the "new" key, > > https://pypi.python.org/serverkey > https://pypi.python.org/serverkey.new > > There will be a key rollover every year on July 1st. At this point, > the "new" key becomes current, and a new "new" key is generated and > published. > > Tools using the mirror infrastructure are encouraged to hard-code > (cache) both keys in their code base, and attempt validation of > a "simple" page against each of them. If both validations fail, > either the keys are outdated, or the mirror they are using has been > compromised. In this case, they should fall back to not using > mirrors, but contact the master directly. > > After a key rollover, mirrors will need some time to refetch the > signatures for all "simple" pages, so the old current key may > be needed some time after the key rollover still. Within one > day, all mirrors should have caught up. > > Also after a key rollover, tools should update their cached copies > of the keys, and produce new releases. Users will then have one > year to upgrade their tools to continue to use the mirror > infrastructure. Tools may also chose to offer users to update > the serverkeys (after due validation), or may chose to download > them automatically, in which case they should validate the https > server certicate for authenticity of the keys. > > The key rollover will be logged in the PyPI journal, > using an empty package name and an empty release. TOOLS USING > THE JOURNAL MAY NEED TO BE FIXED TO ACCOMMODATE EMPTY PACKAGE > NAMES. Earlier today, such a journal entry was already added; > I took it out again when I noticed that some tools actually > do need to be fixed. > > Unless some flaws are detected in this scheme, I'll do the first > key rollover this coming July. > I have another suggestion, what about adding a new response header on https://pypi.python.org/serverkey ? Like "X-Key-Rollover: some date" > Regards, > Martin > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- Tarek Ziad? | http://ziade.org From marcello at perathoner.de Thu Apr 28 13:06:04 2011 From: marcello at perathoner.de (Marcello Perathoner) Date: Thu, 28 Apr 2011 13:06:04 +0200 Subject: [Catalog-sig] The removal of ratings sucks Message-ID: <4DB94A1C.1000707@perathoner.de> I'm upset, because I *was* using the rating system. I also *knew* the limitations of this feature. I didn't think of choosing an application framework based on the rating. I'm not that stupid. (But I have just learned that some people believe that I'm that stupid. Thank you!) The rating system helped me choose the best candidate for quick hacks, but sadly this "battery" is not included anymore. When I'm adding a simple feature that should not take more than 5 minutes to do, I don't want to compare package features, nor see the package's web site, nor see how many tickets are open, nor if the maintainer answers their email ... I want what other users think is the overall best package. I'm surprised that this useful feature has been removed. I'm even more surprised, after reading the relative thread, that the feature has been removed in an effort to kill the messenger by package maintainers that got bad ratings. How could that happen after there was a (pretty fair) poll that showed the majority was wanting a feature like that? Is PyPI run by a vociferous minority? But the most ironic thing in that thread must be that most people, who were against what they said was a "naive" voting system, could not express their opinion better than by saying: +1 In summary: PyPI as a website is now pretty useless for me. I'll just google for "python package doing something" instead. -- Marcello Perathoner webmaster at gutenberg.org From tseaver at palladion.com Thu Apr 28 15:16:30 2011 From: tseaver at palladion.com (Tres Seaver) Date: Thu, 28 Apr 2011 09:16:30 -0400 Subject: [Catalog-sig] The removal of ratings sucks In-Reply-To: <4DB94A1C.1000707@perathoner.de> References: <4DB94A1C.1000707@perathoner.de> Message-ID: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/28/2011 07:06 AM, Marcello Perathoner wrote: > I'm upset, because I *was* using the rating system. I also *knew* the > limitations of this feature. I didn't think of choosing an application > framework based on the rating. I'm not that stupid. (But I have just > learned that some people believe that I'm that stupid. Thank you!) > > The rating system helped me choose the best candidate for quick hacks, > but sadly this "battery" is not included anymore. > > When I'm adding a simple feature that should not take more than 5 > minutes to do, I don't want to compare package features, nor see the > package's web site, nor see how many tickets are open, nor if the > maintainer answers their email ... I want what other users think is the > overall best package. How did you evaluate the care or thought put into the ratings by the tiny handful of PyPI users who actually created them? "Crowd-sourcing" relies on high participation to correct for outliers: without lots of ratings, any one bogus rating, whether derogatory or flattering, carries too much weight. > I'm surprised that this useful feature has been removed. > > I'm even more surprised, after reading the relative thread, that the > feature has been removed in an effort to kill the messenger by package > maintainers that got bad ratings. How could that happen after there was > a (pretty fair) poll that showed the majority was wanting a feature like > that? I don't agree with that characterization of the thread at all: the argument was not "my package got a bad rating, remove ratings", but "quasi-anonymous ratings don't supply enough value". Jacob, for instance, had mostly very high ratings for Django, but had no information on *why* a particular bad rating occurred, and could therefore not address the problem. > Is PyPI run by a vociferous minority? PyPI is run by MvL on behalf of the Python community, with support and feedback (and often not enough gratitude) from the members of this list. In the end, I believe MvL was convinced to remove ratings not by arguments about their usefulness, but by the evidence that almost nobody rated packages. Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver at palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk25aK4ACgkQ+gerLs4ltQ5fLgCeMoxr7S/6QZma7ACL6Kah9l3N UmQAnjG+EhQP7eeX38LCThVTQnhBwypT =KOd3 -----END PGP SIGNATURE----- From martin at v.loewis.de Thu Apr 28 22:03:22 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Thu, 28 Apr 2011 22:03:22 +0200 Subject: [Catalog-sig] PyPI mirror key rollover In-Reply-To: <4DB924C1.600@egenix.com> References: <4DB862E4.9060207@v.loewis.de> <4DB924C1.600@egenix.com> Message-ID: <4DB9C80A.9060806@v.loewis.de> Am 28.04.2011 10:26, schrieb M.-A. Lemburg: > "Martin v. L?wis" wrote: >> I came up with a key rollover scheme for the server key on PyPI. >> [...] >> >> The key rollover will be logged in the PyPI journal, >> using an empty package name and an empty release. TOOLS USING >> THE JOURNAL MAY NEED TO BE FIXED TO ACCOMMODATE EMPTY PACKAGE >> NAMES. Earlier today, such a journal entry was already added; >> I took it out again when I noticed that some tools actually >> do need to be fixed. > > I can't comment on the other parts of the proposal, but the above > suggestions doesn't sound like a good solution: an empty package > name in the update stream looks more like a server or client > decoding bug than a trigger to do a key update. Oops, I forgot a critical detail: the "action" string in the journal entry would be "keyrotate". > Wouldn't it be better to use a descriptive package name such > as "pypi-serverkey-update" together with a package version > which identifies the new serverkey version as trigger ? That would not be good - tools would (rightly) assume that there is a package with that name, and try to mirror it. Regards, Martin From martin at v.loewis.de Thu Apr 28 22:06:32 2011 From: martin at v.loewis.de (=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=) Date: Thu, 28 Apr 2011 22:06:32 +0200 Subject: [Catalog-sig] PyPI mirror key rollover In-Reply-To: References: <4DB862E4.9060207@v.loewis.de> Message-ID: <4DB9C8C8.2010802@v.loewis.de> > I have another suggestion, > what about adding a new response header on https://pypi.python.org/serverkey ? > > Like "X-Key-Rollover: some date" That would be too flakey for several reasons. One reason is that tools may not be able to cope with that format change, either. Plus, I can't really promise that the key rollover happens at this very point. So tools would have to check whether the key rollover actually happened. If they do that, they could as well check whether it happened on every sync, which in turn would be more expensive than necessary. Regards, Martin From mal at egenix.com Thu Apr 28 22:29:29 2011 From: mal at egenix.com (M.-A. Lemburg) Date: Thu, 28 Apr 2011 22:29:29 +0200 Subject: [Catalog-sig] PyPI mirror key rollover In-Reply-To: <4DB9C80A.9060806@v.loewis.de> References: <4DB862E4.9060207@v.loewis.de> <4DB924C1.600@egenix.com> <4DB9C80A.9060806@v.loewis.de> Message-ID: <4DB9CE29.3080609@egenix.com> "Martin v. L?wis" wrote: > Am 28.04.2011 10:26, schrieb M.-A. Lemburg: >> "Martin v. L?wis" wrote: >>> I came up with a key rollover scheme for the server key on PyPI. >>> [...] >>> >>> The key rollover will be logged in the PyPI journal, >>> using an empty package name and an empty release. TOOLS USING >>> THE JOURNAL MAY NEED TO BE FIXED TO ACCOMMODATE EMPTY PACKAGE >>> NAMES. Earlier today, such a journal entry was already added; >>> I took it out again when I noticed that some tools actually >>> do need to be fixed. >> >> I can't comment on the other parts of the proposal, but the above >> suggestions doesn't sound like a good solution: an empty package >> name in the update stream looks more like a server or client >> decoding bug than a trigger to do a key update. > > Oops, I forgot a critical detail: the "action" string in the journal > entry would be "keyrotate". Ah, ok. Makes more sense then :-) >> Wouldn't it be better to use a descriptive package name such >> as "pypi-serverkey-update" together with a package version >> which identifies the new serverkey version as trigger ? > > That would not be good - tools would (rightly) assume that there > is a package with that name, and try to mirror it. Well, you could create a package under that name which then contains a module with all known server keys. Might be useful to have for other purposes as well. -- Marc-Andre Lemburg eGenix.com Professional Python Services directly from the Source (#1, Apr 28 2011) >>> Python/Zope Consulting and Support ... http://www.egenix.com/ >>> mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ >>> mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/ ________________________________________________________________________ 2011-06-20: EuroPython 2011, Florence, Italy 53 days to go ::: 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 http://www.egenix.com/company/contact/ From alexis at notmyidea.org Sat Apr 30 15:22:55 2011 From: alexis at notmyidea.org (=?UTF-8?B?QWxleGlzIE3DqXRhaXJlYXU=?=) Date: Sat, 30 Apr 2011 14:22:55 +0100 Subject: [Catalog-sig] PEP 345 - Environment markers and platform.python_implementation In-Reply-To: <4DBAEBFD.1040508@notmyidea.org> References: <4DBAEBFD.1040508@notmyidea.org> Message-ID: <4DBC0D2F.3060709@notmyidea.org> On 29/04/2011 17:49, Alexis M?taireau wrote: > We recently had a bug report about how to distinguish between different > python platforms (http://bugs.python.org/issue11921) with environment > markers. > > The current version of PEP 345 does not allow to use that > (http://www.python.org/dev/peps/pep-0345/#environment-markers). I'm > willing to add that in packaging/distutils2 and to update the PEP. Is it > alright ? This also means we will need to backport some code when working on packaging ? distutils2 because platform.python_implementation have been introduced in python 2.6. Will update the PEP soon to add that, tell me if you have anything against such change. -- Alexis From ziade.tarek at gmail.com Sat Apr 30 15:38:23 2011 From: ziade.tarek at gmail.com (=?ISO-8859-1?Q?Tarek_Ziad=E9?=) Date: Sat, 30 Apr 2011 15:38:23 +0200 Subject: [Catalog-sig] PEP 345 - Environment markers and platform.python_implementation In-Reply-To: <4DBC0D2F.3060709@notmyidea.org> References: <4DBAEBFD.1040508@notmyidea.org> <4DBC0D2F.3060709@notmyidea.org> Message-ID: 2011/4/30 Alexis M?taireau : > On 29/04/2011 17:49, Alexis M?taireau wrote: >> >> We recently had a bug report about how to distinguish between different >> python platforms (http://bugs.python.org/issue11921) with environment >> markers. >> >> The current version of PEP 345 does not allow to use that >> (http://www.python.org/dev/peps/pep-0345/#environment-markers). I'm >> willing to add that in packaging/distutils2 and to update the PEP. Is it >> alright ? > > This also means we will need to backport some code when working on packaging > ? distutils2 because platform.python_implementation have been introduced in > python 2.6. > > Will update the PEP soon to add that, tell me if you have anything against > such change. I am +1 with the change, but I just need to see if we can do slight changes to the PEP or we need a new PEP will keep you informed. > > -- > Alexis > _______________________________________________ > Catalog-SIG mailing list > Catalog-SIG at python.org > http://mail.python.org/mailman/listinfo/catalog-sig > -- Tarek Ziad? | http://ziade.org