Archiving Threads / Closing Threads
I suggest we find a way to close a mail thread. Obviously we can't stop people from sending mails, but, like when a mod sees the topic has been juiced out, he closes the topic which practically means: if mailman sees mails being sent to a topic closed by a mod, it does not relay it to the list. It is normally difficult to say when a topic has been juiced out (the essentials have been said, with no new inputs). One might close the topic prematuraly. So, a time-based automatic closing might be better. Example: After 5 days, a topic is closed. A mod can extend the closing though. Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius
Mailman, the manager for the mailing list doesn’t offer any such option. The closest that could be done is if a topic just goes on too long, a filter could be added to the list configuration which holds for moderation messages which match the subject. The problem is that if someone changes the subject slightly, it might not get caught by the filter depending on how detailed the filter was.
On Tue, 3 Dec 2019 at 19:52, Richard Damon <Richard@damon-family.org> wrote:
Mailman, the manager for the mailing list doesn’t offer any such option. The closest that could be done is if a topic just goes on too long, a filter could be added to the list configuration which holds for moderation messages which match the subject. The problem is that if someone changes the subject slightly, it might not get caught by the filter depending on how detailed the filter was.
I think it's just the nature of a mailing list. Much as some people might prefer a forum-style approach (and may even be consuming this list through an interface with that sort of style) it's still a mailing list underneath, and that means people can post mails that go to all users essentially without restriction. Imposing restrictions would typically require more work from the mods, who are already overworked. Better to just accept the nature of the communication medium. Paul
Ok posted it with the idea than mailman operated on python and the mods were hackers enough to add functionalities. Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius
https://discuss.python.org is an instance of Discourse, which, like mailman, is also open source. https://github.com/discourse/discourse https://meta.discourse.org/t/discourse-moderation-guide/63116 https://gitlab.com/mailman/mailman I agree that Broken thread reply chains would be the likely outcome of locking mailing list threads: people would frustratedly just create a new thread whereas they would likely just stop contributing to a dead thread. On Tue, Dec 3, 2019, 4:25 PM Abdur-Rahmaan Janhangeer <arj.python@gmail.com> wrote:
Richard Damon writes:
I'm -1 on both moderator-driven closure and time-driven closure (ping me if you care why ;-). That said, Mailman is a Python application and the extension interface is basically "derive a class from Handler or Rule with a method named 'process'." This kind of thing requires very little Mailman expertise (and I assure you the Mailman devs care a lot about Python: if the moderators want you to implement this, we will help). Regarding filtering on the subject, that's not how threading works. If you want to filter on subject as well as close the thread, fine, but the chance that someone accidentally changes the subject and it slips past the filter is very small. Most likely it's a deliberate attempt to either maintain context for a new thread, or evade the filter. Even if the latter, detecting such is still most cheaply done by humans, unless you're Amazon scale (Japanese news analysis today reported that Amazon proactively deletes reviews that its "AI" judges to be inauthentic or generated commercially).
On Dec 3, 2019, at 11:17, Abdur-Rahmaan Janhangeer <arj.python@gmail.com> wrote:
I suggest we find a way to close a mail thread. Obviously we can't stop people from sending mails, but, like when a mod sees the topic has been juiced out, he closes the topic
This does already happen, if only by human process/convention, not technical enforcement. A mod says “We’re just going around in circles here, and it’s getting heated for no good reason, and there’s no actual proposal likely to come out of this, so stop responding unless you have something genuinely worth reopening over.” And I think this has worked well enough without technical enforcement, especially given how rarely it comes up—most threads just die out of their own accord. Occasionally you get a couple more replies over the next few days from people who hadn’t read to the end of the thread, and a filter could block those, but is that a big enough deal to be worth fixing?
It is normally difficult to say when a topic has been juiced out (the essentials have been said, with no new inputs). One might close the topic prematuraly. So, a time-based automatic closing might be better. Example: After 5 days, a topic is closed. A mod can extend the closing though.
Threads that actually end up going somewhere invariably last much longer than 5 days, and requiring a mod to manually extend them (especially repeatedly) would be wasted effort. And, worse, it would be a brand new opportunity to frequently get things wrong, and for people to get used to working around that (which would be trivial to do—just edit the subject), and pretty soon you’ve just got the original problem back but harder to manage than before, plus the threading on many of the actually useful threads is broken. Also, why do you want to encourage such a rush in the first place? The language evolves on 18-month cycles, and changes that aren’t pure library additions often take 3 cycles; if it takes a month to fully discuss an idea, who cares? If some people want to tune out after the first week, they can, and already do. And if someone like Serhiy or even Guido is away for a week, does that mean every thread started during that week should get resolved without their frequently crucial input?
Hi Abdur-Rahmaan, On 4/12/19 8:14 am, Abdur-Rahmaan Janhangeer wrote:
I think a human solution rather than a technical one would be better by either polity asking the one who's going around in circles to do some more useful with their time or by simply ignoring the message. You can also set up filters in your own email program if you feel it's bothering you too much. cheers, Jan
On Tue, Dec 03, 2019 at 11:14:57PM +0400, Abdur-Rahmaan Janhangeer wrote:
What problem are you trying to solve? Are we suffering under a burden of pople resurrecting old threads from ten years ago, or even a year ago? I wish people *would* resurrect old threads, rather than just start a new thread covering the same old ground. At least that way there is a chance that they may have read the previous correspondence. As for your idea of closing threads after five days, *if* we were to do such a thing (and I don't think we should) five days is ludicrously short. It might be okay for people who have nothing better to do than sit at their computer replying to Python-Ideas (and I've been there...) but for people who are busy, or have been sick, or have had computer problems, five days are nothing. I've been in all of those positions: I've been away from home and too busy to read emails for a week, I've been in hospital for two weeks, I've had my ISP silently dropping all email for ten days, etc. The beauty of email is that it is *not* real time: people can respond to threads a week or two later. Don't wreck that feature. -- Steven
Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius On Wed, 4 Dec 2019, 07:38 Steven D'Aprano, <steve@pearwood.info> wrote:
What problem are you trying to solve? Are we suffering under a burden of pople resurrecting old threads from ten years ago, or even a year ago?
Easy digging of threads.
Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius On Wed, 4 Dec 2019, 15:16 Rhodri James, <rhodri@kynesim.co.uk> wrote:
An advantage. If the topic goes round several times, the mods closing the topic freezes the thread so that the thread does not gets filled with unnecessary details. Python threads are bit like a plane from Paris bound for London via direct flight but which sometimes detours through Tokyo and Manilla. In fact you might ask yourself HOW in the world are you in São Paulo. The proposed approached raises some issues such as people opening new threads etc. The idea behind is that list threads are gem mines for Python programmers. If you've followed the thread as it was unspun, that's great. Else if you are reading the archives, you have to untangle the right thread from a mess. Another advantage of the proposed idea might be to make programmers more responsible by writing what is necessary and as in depth as possible.
On 04/12/2019 12:44, Abdur-Rahmaan Janhangeer wrote:
I'm sorry, I misunderstood what you said. I thought that you were complaining that its too easy to dig threads (and misunderstood what "digging threads" meant to you). You're actually complaining about the opposite; that it's too hard to mine threads for information. Sorry, but that's true of any conversation. If you want information out of a thread, you are going to have to put in the work of reading and comprehending it. The only way around that is to get a technical PA to read and comprehend it for you and produce a summary. Sometimes someone is inspired to do that, and a FAQ appears. Mostly you're on your own. Closing down threads isn't going to help. It may reduce the clutter that the solitary piece of information you are interested in is buried under, but it also increases the chance that the information isn't there at all. More to the point, it isn't going to make the thread any less convoluted.
Another advantage of the proposed idea might be to make programmers more responsible by writing what is necessary and as in depth as possible.
Um. I've sat on and chaired a lot of committees in my time, with a lot of different rules of engagement. I have not seen any approach to discussion that has made anyone be more responsible, succinct or accurate. Attempting to curtail discussion usually has the opposite effect, and puts everyone in a bad mood for the next item as well. -- Rhodri James *-* Kynesim Ltd
On Wed, Dec 4, 2019 at 4:47 AM Rhodri James <rhodri@kynesim.co.uk> wrote:
Exactly -- so please contribute to: https://pythonchb.github.io/PythonListsSummaries/python_ideas/index.html The idea is to have a place for thread summaries that don't get to the PEP stage -- there are a LOT iof ideas that get re-discussed here ... -CHB -- Christopher Barker, PhD Python Language Consulting - Teaching - Scientific Software Development - Desktop GUI and Web Development - wxPython, numpy, scipy, Cython
On 12/04/2019 04:44 AM, Abdur-Rahmaan Janhangeer wrote:
An advantage. If the topic goes round several times, the mods closing the topic freezes the thread so that the thread does not gets filled with unnecessary details.
By the time the topic has gone around several times, it is already filled with unnecessary details.
The idea behind is that list threads are gem mines for Python programmers. If you've followed the thread as it was unspun, that's great. Else if you are reading the archives, you have to untangle the right thread from a mess.
It's still a mess even if following it in real-time.
Another advantage of the proposed idea might be to make programmers more responsible by writing what is necessary and as in depth as possible.
I have not seen that happen in /any/ method of discussion. -- ~Ethan~
On Dec 3, 2019, at 19:39, Steven D'Aprano <steve@pearwood.info> wrote:
I agree. And if it were a lot easier to search for past threads on an idea… But StackOverflow couldn’t solve that problem in over a decade of trying, without the limitations of email underneath (you can’t rename the subject of an old thread, link it to a later one, etc.), and all of their quasi-competitors have done even worse, so that’s not a serious proposal, just a blue-sky wish.
Mailman, the manager for the mailing list doesn’t offer any such option. The closest that could be done is if a topic just goes on too long, a filter could be added to the list configuration which holds for moderation messages which match the subject. The problem is that if someone changes the subject slightly, it might not get caught by the filter depending on how detailed the filter was.
On Tue, 3 Dec 2019 at 19:52, Richard Damon <Richard@damon-family.org> wrote:
Mailman, the manager for the mailing list doesn’t offer any such option. The closest that could be done is if a topic just goes on too long, a filter could be added to the list configuration which holds for moderation messages which match the subject. The problem is that if someone changes the subject slightly, it might not get caught by the filter depending on how detailed the filter was.
I think it's just the nature of a mailing list. Much as some people might prefer a forum-style approach (and may even be consuming this list through an interface with that sort of style) it's still a mailing list underneath, and that means people can post mails that go to all users essentially without restriction. Imposing restrictions would typically require more work from the mods, who are already overworked. Better to just accept the nature of the communication medium. Paul
Ok posted it with the idea than mailman operated on python and the mods were hackers enough to add functionalities. Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius
https://discuss.python.org is an instance of Discourse, which, like mailman, is also open source. https://github.com/discourse/discourse https://meta.discourse.org/t/discourse-moderation-guide/63116 https://gitlab.com/mailman/mailman I agree that Broken thread reply chains would be the likely outcome of locking mailing list threads: people would frustratedly just create a new thread whereas they would likely just stop contributing to a dead thread. On Tue, Dec 3, 2019, 4:25 PM Abdur-Rahmaan Janhangeer <arj.python@gmail.com> wrote:
Richard Damon writes:
I'm -1 on both moderator-driven closure and time-driven closure (ping me if you care why ;-). That said, Mailman is a Python application and the extension interface is basically "derive a class from Handler or Rule with a method named 'process'." This kind of thing requires very little Mailman expertise (and I assure you the Mailman devs care a lot about Python: if the moderators want you to implement this, we will help). Regarding filtering on the subject, that's not how threading works. If you want to filter on subject as well as close the thread, fine, but the chance that someone accidentally changes the subject and it slips past the filter is very small. Most likely it's a deliberate attempt to either maintain context for a new thread, or evade the filter. Even if the latter, detecting such is still most cheaply done by humans, unless you're Amazon scale (Japanese news analysis today reported that Amazon proactively deletes reviews that its "AI" judges to be inauthentic or generated commercially).
On Dec 3, 2019, at 11:17, Abdur-Rahmaan Janhangeer <arj.python@gmail.com> wrote:
I suggest we find a way to close a mail thread. Obviously we can't stop people from sending mails, but, like when a mod sees the topic has been juiced out, he closes the topic
This does already happen, if only by human process/convention, not technical enforcement. A mod says “We’re just going around in circles here, and it’s getting heated for no good reason, and there’s no actual proposal likely to come out of this, so stop responding unless you have something genuinely worth reopening over.” And I think this has worked well enough without technical enforcement, especially given how rarely it comes up—most threads just die out of their own accord. Occasionally you get a couple more replies over the next few days from people who hadn’t read to the end of the thread, and a filter could block those, but is that a big enough deal to be worth fixing?
It is normally difficult to say when a topic has been juiced out (the essentials have been said, with no new inputs). One might close the topic prematuraly. So, a time-based automatic closing might be better. Example: After 5 days, a topic is closed. A mod can extend the closing though.
Threads that actually end up going somewhere invariably last much longer than 5 days, and requiring a mod to manually extend them (especially repeatedly) would be wasted effort. And, worse, it would be a brand new opportunity to frequently get things wrong, and for people to get used to working around that (which would be trivial to do—just edit the subject), and pretty soon you’ve just got the original problem back but harder to manage than before, plus the threading on many of the actually useful threads is broken. Also, why do you want to encourage such a rush in the first place? The language evolves on 18-month cycles, and changes that aren’t pure library additions often take 3 cycles; if it takes a month to fully discuss an idea, who cares? If some people want to tune out after the first week, they can, and already do. And if someone like Serhiy or even Guido is away for a week, does that mean every thread started during that week should get resolved without their frequently crucial input?
Hi Abdur-Rahmaan, On 4/12/19 8:14 am, Abdur-Rahmaan Janhangeer wrote:
I think a human solution rather than a technical one would be better by either polity asking the one who's going around in circles to do some more useful with their time or by simply ignoring the message. You can also set up filters in your own email program if you feel it's bothering you too much. cheers, Jan
On Tue, Dec 03, 2019 at 11:14:57PM +0400, Abdur-Rahmaan Janhangeer wrote:
What problem are you trying to solve? Are we suffering under a burden of pople resurrecting old threads from ten years ago, or even a year ago? I wish people *would* resurrect old threads, rather than just start a new thread covering the same old ground. At least that way there is a chance that they may have read the previous correspondence. As for your idea of closing threads after five days, *if* we were to do such a thing (and I don't think we should) five days is ludicrously short. It might be okay for people who have nothing better to do than sit at their computer replying to Python-Ideas (and I've been there...) but for people who are busy, or have been sick, or have had computer problems, five days are nothing. I've been in all of those positions: I've been away from home and too busy to read emails for a week, I've been in hospital for two weeks, I've had my ISP silently dropping all email for ten days, etc. The beauty of email is that it is *not* real time: people can respond to threads a week or two later. Don't wreck that feature. -- Steven
Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius On Wed, 4 Dec 2019, 07:38 Steven D'Aprano, <steve@pearwood.info> wrote:
What problem are you trying to solve? Are we suffering under a burden of pople resurrecting old threads from ten years ago, or even a year ago?
Easy digging of threads.
Abdur-Rahmaan Janhangeer http://www.pythonmembers.club | https://github.com/Abdur-rahmaanJ Mauritius On Wed, 4 Dec 2019, 15:16 Rhodri James, <rhodri@kynesim.co.uk> wrote:
An advantage. If the topic goes round several times, the mods closing the topic freezes the thread so that the thread does not gets filled with unnecessary details. Python threads are bit like a plane from Paris bound for London via direct flight but which sometimes detours through Tokyo and Manilla. In fact you might ask yourself HOW in the world are you in São Paulo. The proposed approached raises some issues such as people opening new threads etc. The idea behind is that list threads are gem mines for Python programmers. If you've followed the thread as it was unspun, that's great. Else if you are reading the archives, you have to untangle the right thread from a mess. Another advantage of the proposed idea might be to make programmers more responsible by writing what is necessary and as in depth as possible.
On 04/12/2019 12:44, Abdur-Rahmaan Janhangeer wrote:
I'm sorry, I misunderstood what you said. I thought that you were complaining that its too easy to dig threads (and misunderstood what "digging threads" meant to you). You're actually complaining about the opposite; that it's too hard to mine threads for information. Sorry, but that's true of any conversation. If you want information out of a thread, you are going to have to put in the work of reading and comprehending it. The only way around that is to get a technical PA to read and comprehend it for you and produce a summary. Sometimes someone is inspired to do that, and a FAQ appears. Mostly you're on your own. Closing down threads isn't going to help. It may reduce the clutter that the solitary piece of information you are interested in is buried under, but it also increases the chance that the information isn't there at all. More to the point, it isn't going to make the thread any less convoluted.
Another advantage of the proposed idea might be to make programmers more responsible by writing what is necessary and as in depth as possible.
Um. I've sat on and chaired a lot of committees in my time, with a lot of different rules of engagement. I have not seen any approach to discussion that has made anyone be more responsible, succinct or accurate. Attempting to curtail discussion usually has the opposite effect, and puts everyone in a bad mood for the next item as well. -- Rhodri James *-* Kynesim Ltd
On Wed, Dec 4, 2019 at 4:47 AM Rhodri James <rhodri@kynesim.co.uk> wrote:
Exactly -- so please contribute to: https://pythonchb.github.io/PythonListsSummaries/python_ideas/index.html The idea is to have a place for thread summaries that don't get to the PEP stage -- there are a LOT iof ideas that get re-discussed here ... -CHB -- Christopher Barker, PhD Python Language Consulting - Teaching - Scientific Software Development - Desktop GUI and Web Development - wxPython, numpy, scipy, Cython
On 12/04/2019 04:44 AM, Abdur-Rahmaan Janhangeer wrote:
An advantage. If the topic goes round several times, the mods closing the topic freezes the thread so that the thread does not gets filled with unnecessary details.
By the time the topic has gone around several times, it is already filled with unnecessary details.
The idea behind is that list threads are gem mines for Python programmers. If you've followed the thread as it was unspun, that's great. Else if you are reading the archives, you have to untangle the right thread from a mess.
It's still a mess even if following it in real-time.
Another advantage of the proposed idea might be to make programmers more responsible by writing what is necessary and as in depth as possible.
I have not seen that happen in /any/ method of discussion. -- ~Ethan~
On Dec 3, 2019, at 19:39, Steven D'Aprano <steve@pearwood.info> wrote:
I agree. And if it were a lot easier to search for past threads on an idea… But StackOverflow couldn’t solve that problem in over a decade of trying, without the limitations of email underneath (you can’t rename the subject of an old thread, link it to a later one, etc.), and all of their quasi-competitors have done even worse, so that’s not a serious proposal, just a blue-sky wish.
participants (11)
-
Abdur-Rahmaan Janhangeer -
Andrew Barnert -
Christopher Barker -
Ethan Furman -
Jan Bakuwel -
Paul Moore -
Rhodri James -
Richard Damon -
Stephen J. Turnbull -
Steven D'Aprano -
Wes Turner