[aur-general] Removing comments from AUR
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
I agree that comments are not terribly useful on community packages, but I think they're important for [unsupported]. Maybe for community packages, only allow comments from the *maintainer* (or TU). My preference is to leave it as is. Granted, sometimes people report issues for community packages on the comment page, but that's no worse than people emailing the maintainer directly (since that's still not on the bug tracker), and is better because it's more visible.
On Wed, Jun 24, 2009 at 4:12 PM, Daenyth Blank<daenyth+arch@gmail.com> wrote:
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
I agree that comments are not terribly useful on community packages, but I think they're important for [unsupported]. Maybe for community packages, only allow comments from the *maintainer* (or TU). My preference is to leave it as is. Granted, sometimes people report issues for community packages on the comment page, but that's no worse than people emailing the maintainer directly (since that's still not on the bug tracker), and is better because it's more visible.
Without sounding harsh - it's not my job. We've always harped on the point that the TUs and community was an autonomous entity. What we are doing here is enabling that even more so. If the TUs and/or AUR guys want to disable comments, that's their prerogative, but there's no way in hell I'm going to add "oh and fiddle with some PHP while we're at it" to the list of things to do here.
On Wed, Jun 24, 2009 at 4:17 PM, Aaron Griffin<aaronmgriffin@gmail.com> wrote:
On Wed, Jun 24, 2009 at 4:12 PM, Daenyth Blank<daenyth+arch@gmail.com> wrote:
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
I agree that comments are not terribly useful on community packages, but I think they're important for [unsupported]. Maybe for community packages, only allow comments from the *maintainer* (or TU). My preference is to leave it as is. Granted, sometimes people report issues for community packages on the comment page, but that's no worse than people emailing the maintainer directly (since that's still not on the bug tracker), and is better because it's more visible.
Without sounding harsh - it's not my job. We've always harped on the point that the TUs and community was an autonomous entity. What we are doing here is enabling that even more so. If the TUs and/or AUR guys want to disable comments, that's their prerogative, but there's no way in hell I'm going to add "oh and fiddle with some PHP while we're at it" to the list of things to do here.
Note, this reply was to Dolby's original comment in the other thread - it looks goofy the way I replied here
Relatedly, can AUR notify you whenever a package of yours gets a new comment? I don't think I get these. -- Andrei Thorp, Developer: Xandros Corp. (http://www.xandros.com) <dark> eat Depends: cook | eat-out. But eat-out is non-free so that's out. And cook Recommends: clean-pans. -- Seen on #Debian
On Thu, Jun 25, 2009 at 12:26 AM, Andrei Thorp <garoth@gmail.com> wrote:
Relatedly, can AUR notify you whenever a package of yours gets a new comment? I don't think I get these. -- Andrei Thorp, Developer: Xandros Corp. (http://www.xandros.com)
<dark> eat Depends: cook | eat-out. But eat-out is non-free so that's out. And cook Recommends: clean-pans. -- Seen on #Debian
Yes it does. You should press the notify button. -- Greg
On Wed, Jun 24, 2009 at 11:26 PM, Andrei Thorp <garoth@gmail.com> wrote:
Relatedly, can AUR notify you whenever a package of yours gets a new comment? I don't think I get these. --
yes AUR can notify you at new comments, there is a checkbox. Point is that the bug tracker is much easier as you can track bugs much easier. In AUR you get notified once and forget about it if you're busy at that moment. Ronald
On Wed, Jun 24, 2009 at 17:36, Ronald van Haren<pressh@gmail.com> wrote:
yes AUR can notify you at new comments, there is a checkbox. Point is that the bug tracker is much easier as you can track bugs much easier. In AUR you get notified once and forget about it if you're busy at that moment.
Ronald
For community packages this is true, but where does that leave packages in unsupported?
On Wed, Jun 24, 2009 at 17:36, Ronald van Haren<pressh@gmail.com> wrote:
yes AUR can notify you at new comments, there is a checkbox. Point is that the bug tracker is much easier as you can track bugs much easier. In AUR you get notified once and forget about it if you're busy at that moment.
Somewhat disagree -- I'd just leave the e-mail in my inbox until I processed it. -- Andrei Thorp, Developer: Xandros Corp. (http://www.xandros.com) NEVER RESPOND TO CRITICAL PRESS. IT IS A GAME YOU CAN ONLY LOSE, AND IT MAKES US LOOK BAD. -- Bruce Perens
On Thu, Jun 25, 2009 at 12:12 AM, Daenyth Blank <daenyth+arch@gmail.com<daenyth%2Barch@gmail.com>
wrote:
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
I agree that comments are not terribly useful on community packages, but I think they're important for [unsupported]. Maybe for community packages, only allow comments from the *maintainer* (or TU). My preference is to leave it as is. Granted, sometimes people report issues for community packages on the comment page, but that's no worse than people emailing the maintainer directly (since that's still not on the bug tracker), and is better because it's more visible.
I disagree on the "not being worst" part. If Sergej who currently maintains more than 40% of the repo on his own read the comments on his packages, i imagine he would spend his life reading comments. See for example the link with brain0's comment. "Reported" as a comment on June 1st, fixed: never reported in the bug tracker on the 16th. fixed on the 19th. And from personal experience i can tell you that he fixes ASAP all the bugs reported to him directly by email. Anyway my suggestion was about unsupported. I thought (but maybe i was mistaking) that you guys (TUs) had already decided to disable comments on community already the last time this was brought up. -- Greg
On Wed, Jun 24, 2009 at 11:28 PM, Grigorios Bouzakis <grbzks@gmail.com>wrote:
On Thu, Jun 25, 2009 at 12:12 AM, Daenyth Blank <daenyth+arch@gmail.com <daenyth%2Barch@gmail.com>< daenyth%2Barch@gmail.com <daenyth%252Barch@gmail.com>>
wrote:
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
Anyway my suggestion was about unsupported. I thought (but maybe i was mistaking) that you guys
I'd rather not have all discussions about AUR packages on this list, I've already quite a few mailing lists to folow and I imagine the way you propose aur-general would become a high-trafic mailing list.
(TUs) had already decided to disable comments on community already the last time this was brought up.
yes I think we did. Ronald
On Thu, Jun 25, 2009 at 12:34 AM, Ronald van Haren <pressh@gmail.com> wrote:
On Wed, Jun 24, 2009 at 11:28 PM, Grigorios Bouzakis <grbzks@gmail.com
wrote:
On Thu, Jun 25, 2009 at 12:12 AM, Daenyth Blank <daenyth+arch@gmail.com <daenyth%2Barch@gmail.com> < daenyth%2Barch@gmail.com <daenyth%252Barch@gmail.com>>< daenyth%2Barch@gmail.com <daenyth%252Barch@gmail.com> < daenyth%252Barch@gmail.com <daenyth%25252Barch@gmail.com>>>
wrote:
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
Anyway my suggestion was about unsupported. I thought (but maybe i was mistaking) that you guys
I'd rather not have all discussions about AUR packages on this list, I've already quite a few mailing lists to folow and I imagine the way you propose aur-general would become a high-trafic mailing list.
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes. -- Greg
On Wed, Jun 24, 2009 at 17:43, Grigorios Bouzakis<grbzks@gmail.com> wrote:
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes.
-- Greg
I would be alright with this... Perhaps a setting like controllable by the maintainer that says "allow comments"?
On Thu, Jun 25, 2009 at 12:45 AM, Daenyth Blank<daenyth+arch@gmail.com> wrote:
On Wed, Jun 24, 2009 at 17:43, Grigorios Bouzakis<grbzks@gmail.com> wrote:
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes.
-- Greg
I would be alright with this... Perhaps a setting like controllable by the maintainer that says "allow comments"?
I have no problem with extra options, but i imagine that this would increase complexity in the code. I hear a patches welcome coming my way :P Anyway i think both of my suggestions are easy to implemend and viable. -- Greg
On Wed, Jun 24, 2009 at 11:45 PM, Daenyth Blank <daenyth+arch@gmail.com<daenyth%2Barch@gmail.com>
wrote:
On Wed, Jun 24, 2009 at 17:43, Grigorios Bouzakis<grbzks@gmail.com> wrote:
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes.
-- Greg
I would be alright with this... Perhaps a setting like controllable by the maintainer that says "allow comments"?
why not allow the maintainers in unsupported to delete comments for their packages, I don't think it will be too much misused? I remove from time to time the crap out of the comments in my community/aur packages so only the more relevant things stay (if there are any). On Wed, Jun 24, 2009 at 11:41 PM, Daenyth Blank <daenyth+arch@gmail.com<daenyth%2Barch@gmail.com>
wrote:
yes AUR can notify you at new comments, there is a checkbox. Point is
On Wed, Jun 24, 2009 at 17:36, Ronald van Haren<pressh@gmail.com> wrote: that
the bug tracker is much easier as you can track bugs much easier. In AUR you get notified once and forget about it if you're busy at that moment.
Ronald
For community packages this is true, but where does that leave packages in unsupported?
I don't want to remove comments from them if that is what you mean. See my comment above. Ronald
On Wed 24 Jun 2009 23:53 +0200, Ronald van Haren wrote:
why not allow the maintainers in unsupported to delete comments for their packages, I don't think it will be too much misused? I remove from time to time the crap out of the comments in my community/aur packages so only the more relevant things stay (if there are any).
Yeah I think this idea is best out of the lot.
On Wed, Jun 24, 2009 at 18:58, Loui Chang<louipc.ist@gmail.com> wrote:
On Wed 24 Jun 2009 23:53 +0200, Ronald van Haren wrote:
why not allow the maintainers in unsupported to delete comments for their packages, I don't think it will be too much misused? I remove from time to time the crap out of the comments in my community/aur packages so only the more relevant things stay (if there are any).
Yeah I think this idea is best out of the lot.
+1 from me.
On Thu, Jun 25, 2009 at 1:58 AM, Loui Chang <louipc.ist@gmail.com> wrote:
why not allow the maintainers in unsupported to delete comments for their packages, I don't think it will be too much misused? I remove from time to time the crap out of the comments in my community/aur packages so only
On Wed 24 Jun 2009 23:53 +0200, Ronald van Haren wrote: the
more relevant things stay (if there are any).
Yeah I think this idea is best out of the lot.
Yeah it is a nice idea, only i dont think people are gonna spend time deleting comments. Usually when a somewhat popular package gets uploaded theres a lot of discussion about how to do this & that etc. Then when the script reaches a fairly satisfying point of decency discussion stops. & then you mostly see comments like "1.2 is out" or complete build scripts or patches posted etc. See for example http://aur.archlinux.org/packages.php?ID=24266 IMO its better that the first, and more important part of the discussion to be done on the mailing list, this way people who dont use the package can help. People can also get familiar with the package that way. Then email the maintainer directly. -- Greg
2009/6/25 Grigorios Bouzakis <grbzks@gmail.com>:
On Thu, Jun 25, 2009 at 1:58 AM, Loui Chang <louipc.ist@gmail.com> wrote:
why not allow the maintainers in unsupported to delete comments for their packages, I don't think it will be too much misused? I remove from time to time the crap out of the comments in my community/aur packages so only
On Wed 24 Jun 2009 23:53 +0200, Ronald van Haren wrote: the
more relevant things stay (if there are any).
Yeah I think this idea is best out of the lot.
Yeah it is a nice idea, only i dont think people are gonna spend time deleting comments.
Usually when a somewhat popular package gets uploaded theres a lot of discussion about how to do this & that etc. Then when the script reaches a fairly satisfying point of decency discussion stops. & then you mostly see comments like "1.2 is out" or complete build scripts or patches posted etc. See for example http://aur.archlinux.org/packages.php?ID=24266
IMO its better that the first, and more important part of the discussion to be done on the mailing list, this way people who dont use the package can help. People can also get familiar with the package that way. Then email the maintainer directly.
Hi, I think there are way larger number of people maintaining packages in the "unsupported" than how many people are on this mailing list. In "unsupported" I don't agree with removing comments, or even deleting them for that matter. This is because the comments provide invaluable resource for information in one place. Adopted a neglected package that is way out of date? Look among the comments and more likely than not someone already did some modifications that made it work but didn't want to adopt the package. That would be a pain to find (even if you want to look for it) in a mailing list. And if the mail goes only to the developer, then it's all gone when they ignore it. Or bugs? How many "unsupported" maintainer would read the bug tracker to look for their own packages if someone filed a bug? Place a comment, and it's in one place, even get a notification. There's a reason why packages are in "unsupported" and not higher up.... I think the current "notify" and "flag" buttons are adequate for the most "hobby maintainers". For "community", that's of course a different matter, just get rid of comments and force people to file bugs. Maybe that will improve the standards in the "unsupported" as well..... Cheers, Greg
On Thu, Jun 25, 2009 at 6:09 AM, Gergely Imreh <imrehg@gmail.com> wrote:
On Thu, Jun 25, 2009 at 1:58 AM, Loui Chang <louipc.ist@gmail.com> wrote:
On Wed 24 Jun 2009 23:53 +0200, Ronald van Haren wrote:
why not allow the maintainers in unsupported to delete comments for
2009/6/25 Grigorios Bouzakis <grbzks@gmail.com>: their
packages, I don't think it will be too much misused? I remove from time to time the crap out of the comments in my community/aur packages so only the more relevant things stay (if there are any).
Yeah I think this idea is best out of the lot.
Yeah it is a nice idea, only i dont think people are gonna spend time deleting comments.
Usually when a somewhat popular package gets uploaded theres a lot of discussion about how to do this & that etc. Then when the script reaches a fairly satisfying point of decency discussion stops. & then you mostly see comments like "1.2 is out" or complete build scripts or patches posted etc. See for example http://aur.archlinux.org/packages.php?ID=24266
IMO its better that the first, and more important part of the discussion to be done on the mailing list, this way people who dont use the package can help. People can also get familiar with the package that way. Then email the maintainer directly.
Hi,
I think there are way larger number of people maintaining packages in the "unsupported" than how many people are on this mailing list.
Thats true. The discussion could also be moved to the bbs, I guess more people visit that.
In "unsupported" I don't agree with removing comments, or even deleting them for that matter. This is because the comments provide invaluable resource for information in one place. Adopted a neglected package that is way out of date? Look among the comments and more likely than not someone already did some modifications that made it work but didn't want to adopt the package. That would be a pain to find (even if you want to look for it) in a mailing list. And if the mail goes only to the developer, then it's all gone when they ignore it.
Then he/she should adopt the package and fix the script. Or contact the maintainer to fix it instead. Not post a comment. The AUR is not of forum, but in many cases it resembles one.
Or bugs? How many "unsupported" maintainer would read the bug tracker to look for their own packages if someone filed a bug? Place a comment, and it's in one place, even get a notification.
I am not suggesting that.
There's a reason why packages are in "unsupported" and not higher up.... I think the current "notify" and "flag" buttons are adequate for the most "hobby maintainers".
I dont know what reason you are reffering to, i can only guess you are talking about the quality of the scripts. Well for what its worth, application developers seem to be agree with you. But dont forget that unsupported currecntly has 15000 scripts, Lets say that 10000 of them are unique. That is still 2,5 times more than the size of all binary repositories together.
For "community", that's of course a different matter, just get rid of comments and force people to file bugs. Maybe that will improve the standards in the "unsupported" as well.....
Cheers, Greg
-- Greg
I have to agree that the AUR comments "framework" is looking more and more like a forum..In view of that, it meets the criteria to be scrapped entirely. As far as I can tell, there is already a forum section for AUR packages, and discussions can be initiated there if the need arises. A forum thread may be linked in the web interface, so here we may have to slot in some extra PHP. This would be akin to how the maintainer in unsupported can set the category (which should be scrapped too).
Hi, I'm one of those "hobby maintainers", who maintain only a few packages on AUR. And I agree with Gergely Imreh. Moving the AUR comments to the mailing list or to the bbs is not an option. I'm only on this mailing list, because a while ago I had an orphan request. I guess most maintainers like me are not on this mailing list. And finding comments on this mailing list, e.g. in the archive, is not really funny. I think many comments wouldn't be noticed by the appropriate maintainer. So moving the comments to the mailing list is definitely not an option. Btw., the mailing list is called "aur-general", so I think, that this mailing list is about AUR itself rather than single AUR packages. Moving the comments to the bbs is also not an option, because I don't look into the bbs every day. I usually only look into the bbs, if I've got a problem, which is not a bug, and need to find a solution. If the AUR comments would be moved to the bbs, I had to look into the bbs nearly every day, only because it could be, that there is a new thread regarding my packages. This is not practicable and too time consuming. Sending mails to the maintainer directly is principally possible, but if a maintainer of a package is not reachable, either temporarily or permanently (mail server is down, e-mail account doesn't exist anymore, he switches to a new ISP, or whatever), those comments go to Nirvana. If a maintainer gets an issue by e-mail, but doesn't maintain the package anymore, doesn't orphan the package, and doesn't respond to such comments, which is unfortunately not so unusual, the comments are gone and a new maintainer can't work on these issues. In the AUR I can set the notify option and get a mail as soon as a comment regarding one of my packages is posted, I know at once, which package is affected, and I can look at this comment at once. Also this way I can respond much faster. And I don't have to look every day for new possible comments. Also the users, who are interested in an AUR package can see, if there are problems with a package, and can decide not to install a package. This also increases the security of the AUR packages, because if someone finds a security issue (see the "rm -f /*" example), he can send a warning to the comments, so that every user can see it at first glance. I guess not every user can read and understand shell scripts. And I don't see, that the comments are misused as a forum in most cases. Usually there are only comments regarding bugs, feature request, etc. And for the cases, in which the AUR comments are misused as a forum, there could be the possibility for the maintainer to delete such comments. With such a delete function the maintainer could also delete obsolete comments. Another point is, that the maintainer can give important notes about his package, which need to be known before a package gets installed. See e.g. my comments regarding the renaming of the fbcondecor kernel from kernel26fbcondecor to kernel26-fbcondecor, or the hint, that two specific graphic cards, which are supported by kernel26 are not supported by kernel26-fbcondecor, etc. I'm not sure, if it's useful to use the bug tracker for AUR packages, because I haven't used a bug tracker as a developer yet. If I have to look into the bug tracker every day to see, if there is a possible bug report for my package, then I don't think that this is an option, because I don't get so many comments for my packages. I agree, that an issue, which needs a longer discussion, more testing or whatever should be moved from the AUR comments to private e-mails with the maintainer directly. But I haven't seen many of those cases in the AUR comments. So I don't see an alternative for the AUR comments, but I also vote for a delete function for the maintainer. Cheers, Heiko
Am Thu, 25 Jun 2009 10:28:53 +0200 schrieb Heiko Baums <lists@baums-on-web.de>:
I guess most maintainers like me are not on this mailing list. And finding comments on this mailing list, e.g. in the archive, is not really funny. I think many comments wouldn't be noticed by the appropriate maintainer. So moving the comments to the mailing list is definitely not an option.
Forgot to mention, that not every user, who wants to install one or two packages from AUR, is on this mailing list, knows about this mailing list, and wants to subscribe to a mailing list just to be able to post one comment. Cheers, Heiko
I agree with Heiko regarding the utility of the comments for unsupported packages. They provide a single location for posting caveats, updated PKGBUILDs, discussing the particulars of building the PKG, etc. It would be sheer folly to remove them and expect users and maintainers to hunt down this information elsewhere. The current system works and it works well. Xyne
Hi,
I'm one of those "hobby maintainers", who maintain only a few packages on AUR.
And I agree with Gergely Imreh. Moving the AUR comments to the mailing list or to the bbs is not an option.
I'm only on this mailing list, because a while ago I had an orphan request.
I guess most maintainers like me are not on this mailing list. And finding comments on this mailing list, e.g. in the archive, is not really funny. I think many comments wouldn't be noticed by the appropriate maintainer. So moving the comments to the mailing list is definitely not an option.
Btw., the mailing list is called "aur-general", so I think, that this mailing list is about AUR itself rather than single AUR packages.
Moving the comments to the bbs is also not an option, because I don't look into the bbs every day. I usually only look into the bbs, if I've got a problem, which is not a bug, and need to find a solution. If the AUR comments would be moved to the bbs, I had to look into the bbs nearly every day, only because it could be, that there is a new thread regarding my packages. This is not practicable and too time consuming.
Sending mails to the maintainer directly is principally possible, but if a maintainer of a package is not reachable, either temporarily or permanently (mail server is down, e-mail account doesn't exist anymore, he switches to a new ISP, or whatever), those comments go to Nirvana. If a maintainer gets an issue by e-mail, but doesn't maintain the package anymore, doesn't orphan the package, and doesn't respond to such comments, which is unfortunately not so unusual, the comments are gone and a new maintainer can't work on these issues.
In the AUR I can set the notify option and get a mail as soon as a comment regarding one of my packages is posted, I know at once, which package is affected, and I can look at this comment at once. Also this way I can respond much faster. And I don't have to look every day for new possible comments.
Also the users, who are interested in an AUR package can see, if there are problems with a package, and can decide not to install a package. This also increases the security of the AUR packages, because if someone finds a security issue (see the "rm -f /*" example), he can send a warning to the comments, so that every user can see it at first glance. I guess not every user can read and understand shell scripts.
And I don't see, that the comments are misused as a forum in most cases. Usually there are only comments regarding bugs, feature request, etc. And for the cases, in which the AUR comments are misused as a forum, there could be the possibility for the maintainer to delete such comments.
With such a delete function the maintainer could also delete obsolete comments.
Another point is, that the maintainer can give important notes about his package, which need to be known before a package gets installed. See e.g. my comments regarding the renaming of the fbcondecor kernel from kernel26fbcondecor to kernel26-fbcondecor, or the hint, that two specific graphic cards, which are supported by kernel26 are not supported by kernel26-fbcondecor, etc.
I'm not sure, if it's useful to use the bug tracker for AUR packages, because I haven't used a bug tracker as a developer yet. If I have to look into the bug tracker every day to see, if there is a possible bug report for my package, then I don't think that this is an option, because I don't get so many comments for my packages.
I agree, that an issue, which needs a longer discussion, more testing or whatever should be moved from the AUR comments to private e-mails with the maintainer directly. But I haven't seen many of those cases in the AUR comments.
So I don't see an alternative for the AUR comments, but I also vote for a delete function for the maintainer.
Cheers, Heiko
On Thu 25 Jun 2009 00:43 +0300, Grigorios Bouzakis wrote:
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes.
So you're asking that comments be changed to 'Report a bug?' or what? That's what they're used for 90% of the time in my experience. For bug reporting, and follow up and "It works now, thanks for fixing."
On Thu, Jun 25, 2009 at 12:53 AM, Loui Chang <louipc.ist@gmail.com> wrote:
On Thu 25 Jun 2009 00:43 +0300, Grigorios Bouzakis wrote:
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes.
So you're asking that comments be changed to 'Report a bug?' or what?
No, the way i thought of this was for example a box like the one saying "Enter your comment below." that instead says "Package Notes. -- Greg
On Thu, Jun 25, 2009 at 12:58 AM, Grigorios Bouzakis <grbzks@gmail.com>wrote:
On Thu, Jun 25, 2009 at 12:53 AM, Loui Chang <louipc.ist@gmail.com> wrote:
On Thu 25 Jun 2009 00:43 +0300, Grigorios Bouzakis wrote:
Maybe the maintainer should be able to add comments then? Instead of having 100 user comments out of which 5-10, at best, are relevant and useful. eg. One package needs -Sf cause its a custom kernel or something I email the maintainer and he adds it to his notes.
So you're asking that comments be changed to 'Report a bug?' or what?
No, the way i thought of this was for example a box like the one saying "Enter your comment below." that instead says "Package Notes.
Sorry i sent this by accident. This would replace the comment box and it would be editable only by the maintainer. The user would still have to email the maintainer for any issues. -- Greg
On Wed, Jun 24, 2009 at 17:03, Grigorios Bouzakis<grbzks@gmail.com> wrote:
May i suggest something slightly relevant, but probably very radical? Disable comments on AUR completely.
I don't understand why this was even proposed. It seems rather obvious to me that this is not a wise idea: how is a package maintainer to know if someone opened a thread on his package on the bbs? Does every maintainer get his own personal forum? Same goes for the ml: imagine the whole AUR comments becoming e-mails... it would be chaotic to say the least. Forums and mailing lists are general-purpose instruments, and for this very nature they can't be very structured. They are also open to abuse. The AUR has a specialized function, and it does it well. Why split it into two? Keep the information where it belongs (on the package page) and keep a clean structure (don't put everything in one place without easy ways to filter). Where's the KISS philosophy? This is a BIG -1 from me. Corrado
On Thu, Jun 25, 2009 at 10:44 AM, bardo<ilbardo@gmail.com> wrote:
Keep the information where it belongs (on the package page) and keep a clean structure (don't put everything in one place without easy ways to filter). Where's the KISS philosophy?
This is a BIG -1 from me.
Having the information on the package page is indeed very practical. I guess the problem is that there is absolutely no organization and structure of that information. No way to group messages by problems, to easily see which points are still open / relevant, etc.
2009/6/25 Xavier <shiningxc@gmail.com>:
Having the information on the package page is indeed very practical.
I guess the problem is that there is absolutely no organization and structure of that information. No way to group messages by problems, to easily see which points are still open / relevant, etc.
If you intend to propose a bug tracker for the AUR, I think this is overkill :) I think it's good this way... even though it could be improved. Threaded comments? Requests for comment deletion to the package owner? This is what comes to the top of my mind. Corrado
On Thu, Jun 25, 2009 at 12:36 PM, Xavier <shiningxc@gmail.com> wrote:
On Thu, Jun 25, 2009 at 10:44 AM, bardo<ilbardo@gmail.com> wrote:
Keep the information where it belongs (on the package page) and keep a clean structure (don't put everything in one place without easy ways to filter). Where's the KISS philosophy?
This is a BIG -1 from me.
Having the information on the package page is indeed very practical.
I guess the problem is that there is absolutely no organization and structure of that information. No way to group messages by problems, to easily see which points are still open / relevant, etc.
Thats why i suggested having an array where the maintainer can post notes instead of being able to add comments. I opened the last 10 packages updated/uploaded to the AUR & read the comments. http://aur.archlinux.org/packages.php?ID=15207 http://aur.archlinux.org/packages.php?ID=25042 (no comments yet) http://aur.archlinux.org/packages.php?ID=15040 http://aur.archlinux.org/packages.php?ID=6839 http://aur.archlinux.org/packages.php?ID=19209 http://aur.archlinux.org/packages.php?ID=12491 http://aur.archlinux.org/packages.php?ID=22740 (no comments yet) http://aur.archlinux.org/packages.php?ID=27610 (no comments yet) http://aur.archlinux.org/packages.php?ID=18994 http://aur.archlinux.org/packages.php?ID=19634 Out of all the comments i read, close to 100, i found 1 that could be useful to the Arch user. And that is that snes9x-gtk is on the Arch games repo. That could be added to the package notes. All others ones interest only the maintainer. Those could be addressed to him personally. PS. Pacman developers worked hard to add Changelog support. Still a comment is easier than including a file. Conviniency seems to be the root of all evil. PS2. I never said that this wouldnt change the way things are now. I said the exact opposite. -- Greg
Am Thu, 25 Jun 2009 13:23:36 +0300 schrieb Grigorios Bouzakis <grbzks@gmail.com>:
This is an example, why sending comments to the maintainer directly is not the best idea. The original contributor hasn't fixed the issues and had no interest in maintaining it anymore. Now the package has a new maintainer. If the comments had been posted to the maintainer directly by e-mail, then the new maintainer wouldn't know anything about these issues and wouldn't be able to fix them, even if it seems, that he hasn't done it anyway, but that's a different story. Cheers, Heiko
On Thu, Jun 25, 2009 at 1:44 PM, Heiko Baums<lists@baums-on-web.de> wrote:
Am Thu, 25 Jun 2009 13:23:36 +0300 schrieb Grigorios Bouzakis <grbzks@gmail.com>:
This is an example, why sending comments to the maintainer directly is not the best idea.
The original contributor hasn't fixed the issues and had no interest in maintaining it anymore. Now the package has a new maintainer. If the comments had been posted to the maintainer directly by e-mail, then the new maintainer wouldn't know anything about these issues and wouldn't be able to fix them, even if it seems, that he hasn't done it anyway, but that's a different story.
Cheers, Heiko
Thanks for using an example. I also think its a good one to prove my point. IMO you are wrong. And i will explain why. Imagine theres no comments. The user who wrote the second comment, i guess you are reffering to him, would have contacted the maintainer with his suggestion. The maintainer would have either a) implemented it b) told him he didnt like it & wont implemend it. c) told him he doesnt maintain the package anymore. d) wouldnt respond. For the first two, i dont have to explain further. On c and d, it would be up to cyberpatrol who is obviously interested in the package to send for example an email on this mailing list and claim the ownership of the package. Instead, what happened. The package was left unmaintained for 1.5 year (i assume it was updated today which might be wrong) but anyway it was certainly left unmaintained for some time , and the issue that this comment was meant to fix, is still not fixed. -- Greg
On Thu 25 Jun 2009 14:03 +0300, Grigorios Bouzakis wrote:
Instead, what happened. The package was left unmaintained for 1.5 year (i assume it was updated today which might be wrong) but anyway it was certainly left unmaintained for some time , and the issue that this comment was meant to fix, is still not fixed.
Reason: No one really wants to maintain it. Not because AUR has comments.
When I first started submitting to AUR, I was surprised that there was no option to delete comments. My second thought was "Oh well, they are probably deleted automatically after a certain amount of time". Having the comments just linger on, without being able to weed out irrelevant comments seems like a bad idea. How about being able to "pin" comments, by pressing a little icon of a pin, preventing the comment from being deleted, while all other comments are removed automatically when they are one month old? I think this would clean up the comments, while keeping only the relevant ones. Cheers, Alexander
On Thu 25 Jun 2009 13:23 +0300, Grigorios Bouzakis wrote:
On Thu, Jun 25, 2009 at 12:36 PM, Xavier <shiningxc@gmail.com> wrote:
On Thu, Jun 25, 2009 at 10:44 AM, bardo<ilbardo@gmail.com> wrote:
Keep the information where it belongs (on the package page) and keep a clean structure (don't put everything in one place without easy ways to filter). Where's the KISS philosophy?
This is a BIG -1 from me.
Having the information on the package page is indeed very practical.
I guess the problem is that there is absolutely no organization and structure of that information. No way to group messages by problems, to easily see which points are still open / relevant, etc.
Thats why i suggested having an array where the maintainer can post notes instead of being able to add comments.
I opened the last 10 packages updated/uploaded to the AUR & read the comments. http://aur.archlinux.org/packages.php?ID=15207 http://aur.archlinux.org/packages.php?ID=25042 (no comments yet) http://aur.archlinux.org/packages.php?ID=15040 http://aur.archlinux.org/packages.php?ID=6839 http://aur.archlinux.org/packages.php?ID=19209 http://aur.archlinux.org/packages.php?ID=12491 http://aur.archlinux.org/packages.php?ID=22740 (no comments yet) http://aur.archlinux.org/packages.php?ID=27610 (no comments yet) http://aur.archlinux.org/packages.php?ID=18994 http://aur.archlinux.org/packages.php?ID=19634
Out of all the comments i read, close to 100, i found 1 that could be useful to the Arch user. And that is that snes9x-gtk is on the Arch games repo. That could be added to the package notes.
All others ones interest only the maintainer. Those could be addressed to him personally.
PS. Pacman developers worked hard to add Changelog support. Still a comment is easier than including a file. Conviniency seems to be the root of all evil.
PS2. I never said that this wouldnt change the way things are now. I said the exact opposite.
Alright. Removing comments is not the solution, but a change in how they work, how they're organised, and how users are notified about changes in packages may be in order. Specific ideas on how that can be done are always welcome in the AUR bug tracker. Cheers!
On Thu, Jun 25, 2009 at 2:07 PM, Loui Chang<louipc.ist@gmail.com> wrote:
Alright. Removing comments is not the solution, but a change in how they work, how they're organised, and how users are notified about changes in packages may be in order.
Specific ideas on how that can be done are always welcome in the AUR bug tracker.
Cheers!
Since i started this, even by stupidly replying to another thread, i might as well answer to that. My suggestion is not having comments in the AUR at all, comments arent useful to the users. They are only useful to the maintainer. If there is no maintainer, if a user feels the need to comment with an updated/altered script then he should adopt it and fix it. And even disown it afterwards if he feels like it. Yes this will spoil your conviniency of just leaving a comment and pretend you care. Yes this will change the way you are used to doing things. My suggestion is to remove the notify button and the comments. Add a box where the maintainer can add the notes that he feels might be useful to the users. Discussion regarding the way the script builds the package, bugs, etc should be done between the interested party and the maintainer. I dont see the need to add anything special to handle that. Like a "contact maintainer button". If the user is too lazy to follow the maintainer link and then the link to his email, then he shouldnt be motivated enough to comment either. If there is need for further discussion, the script is horribly outdated and unmaintained, should be done, like its done now, on the mailing list or the bbs. How's that for "KISS" ? -- Greg
On Fri, Jun 26, 2009 at 3:23 AM, Grigorios Bouzakis<grbzks@gmail.com> wrote:
On Thu, Jun 25, 2009 at 2:07 PM, Loui Chang<louipc.ist@gmail.com> wrote:
Alright. Removing comments is not the solution, but a change in how they work, how they're organised, and how users are notified about changes in packages may be in order.
Specific ideas on how that can be done are always welcome in the AUR bug tracker.
Cheers!
Since i started this, even by stupidly replying to another thread, i might as well answer to that. My suggestion is not having comments in the AUR at all, comments arent useful to the users. They are only useful to the maintainer. If there is no maintainer, if a user feels the need to comment with an updated/altered script then he should adopt it and fix it. And even disown it afterwards if he feels like it.
Yes this will spoil your conviniency of just leaving a comment and pretend you care. Yes this will change the way you are used to doing things.
My suggestion is to remove the notify button and the comments. Add a box where the maintainer can add the notes that he feels might be useful to the users. Discussion regarding the way the script builds the package, bugs, etc should be done between the interested party and the maintainer. I dont see the need to add anything special to handle that. Like a "contact maintainer button". If the user is too lazy to follow the maintainer link and then the link to his email, then he shouldnt be motivated enough to comment either. If there is need for further discussion, the script is horribly outdated and unmaintained, should be done, like its done now, on the mailing list or the bbs.
How's that for "KISS" ?
Am Fri, 26 Jun 2009 03:23:22 +0300 schrieb Grigorios Bouzakis <grbzks@gmail.com>:
How's that for "KISS" ?
KISS are comments for the packages. bbs, mailing list, bug tracker and mails directly to the maintainer are the opposite of KISS in my opinion. And comments are not only useful for the maintainers but also for the users. Cheers, Heiko
Since i started this, even by stupidly replying to another thread, i might as well answer to that. My suggestion is not having comments in the AUR at all, comments arent useful to the users. They are only useful to the maintainer.
I would disagree... Sometimes it is good for maintaner-user feedback as well. E.g. one of my packages takes a long time to compile. It's a small package but one step looks as if it hung and stays there for about 10-15mins on my computer. I had one of the users place a comment that his compilation didn't work, it froze and he had to kill it after about 5 minutes. Told him to wait a bit longer and it worked. Sure it could be done in the BBS - but would be completely inefficient. Not many users for most of the packages so not many people know what's going on, and I'm not going to search through the forums every day to see if someone wrote about them... Now: comment placed, me notified, can act on it.... Or: someone makes a package. A TU or a more knowledgeable user points out some problems with it in comments so he can fix it. Other users see the comments and see the advice, one day when they will make packages they can take that advice that is now public, and not in someone's mailbox. Yeah, the Wiki is for such things, but how many little things are there that people don't Wiki up? Or how many time people still write on the mailing list while this and that does not work in a package when it should? Sure it "only" concerns the maintainer at that time but there are a much wider potential audience. Also, it can serve as a "call" for other users who are interested in that package (and probably set it to "notify") to call for someone else to adopt a package. And also, your assumption is 100% reliable dedicated knowledgeable maintainers. Which is obviously not the case...
If there is no maintainer, if a user feels the need to comment with an updated/altered script then he should adopt it and fix it. And even disown it afterwards if he feels like it.
Not everyone is the same. Not everyone wants to take that responsibility. Is forcing them the right way? I'd see more people giving up a package before adopting it. If they want to adopt they would do it under the current arrangement. Though the "email the list if one package is very outdated and the maintainer don't give a" is probably not that clear for everyone, might be better to advertise it a bit more, but that's a different issue.
How's that for "KISS" ?
I'd ask, if you have a website that supposed to be automated and self contained, how is it KISS to require people +1 registration to BBS, +1 registration to mailing list, +possibly much longer waiting to contact someone who can know the solution to the problem? Just thinking... Greg
On Fri, Jun 26, 2009 at 3:42 AM, Gergely Imreh<imrehg@gmail.com> wrote:
Since i started this, even by stupidly replying to another thread, i might as well answer to that. My suggestion is not having comments in the AUR at all, comments arent useful to the users. They are only useful to the maintainer.
I would disagree... Sometimes it is good for maintaner-user feedback as well. E.g. one of my packages takes a long time to compile. It's a small package but one step looks as if it hung and stays there for about 10-15mins on my computer. I had one of the users place a comment that his compilation didn't work, it froze and he had to kill it after about 5 minutes. Told him to wait a bit longer and it worked. Sure it could be done in the BBS - but would be completely inefficient. Not many users for most of the packages so not many people know what's going on, and I'm not going to search through the forums every day to see if someone wrote about them... Now: comment placed, me notified, can act on it....
Scrap the bbs. If he mailed you, wouldnt you have bothered to reply his email? Or it needs to be done by a comment or not done at all? You wouldnt have added it to the package notes because it wouldnt interest the users? If you had, he wouldnt even have to email you.
Or: someone makes a package. A TU or a more knowledgeable user points out some problems with it in comments so he can fix it. Other users see the comments and see the advice, one day when they will make packages they can take that advice that is now public, and not in someone's mailbox. Yeah, the Wiki is for such things, but how many little things are there that people don't Wiki up? Or how many time people still write on the mailing list while this and that does not work in a package when it should? Sure it "only" concerns the maintainer at that time but there are a much wider potential audience.
Again private email. bbs & mailing list & IRC = last resort.
Also, it can serve as a "call" for other users who are interested in that package (and probably set it to "notify") to call for someone else to adopt a package.
mailing list, bbs, IRC
And also, your assumption is 100% reliable dedicated knowledgeable maintainers. Which is obviously not the case...
mailing list, bbs, IRC
If there is no maintainer, if a user feels the need to comment with an updated/altered script then he should adopt it and fix it. And even disown it afterwards if he feels like it.
Not everyone is the same. Not everyone wants to take that responsibility. Is forcing them the right way? I'd see more people giving up a package before adopting it. If they want to adopt they would do it under the current arrangement. Though the "email the list if one package is very outdated and the maintainer don't give a" is probably not that clear for everyone, might be better to advertise it a bit more, but that's a different issue.
Then dont. What is the purpose of adding an updated PKGBUILD as a comment? An irresponsible way to update the package?
How's that for "KISS" ?
I'd ask, if you have a website that supposed to be automated and self contained, how is it KISS to require people +1 registration to BBS, +1 registration to mailing list, +possibly much longer waiting to contact someone who can know the solution to the problem?
Just thinking... Greg
Those channels are already set up and working reliably. You just have to use them properly. Just like its done with the binary repositories. Do you mean the way the official Arch works is wrong? Its not KISS? If you are not willing to get involved you shouldnt "maintain" AUR scripts. PS. Please dont overexpand on the KISS issue. I just wrote that because for some strange reason it freequently gets mentioned as an aliby not to fix obvious problems. Its totally irrelevant. KISS doesnt neither say that all pieces of unseful information should be gathered in one place and maintain a chaos. -- Greg
On Fri, Jun 26, 2009 at 8:33 PM, Grigorios Bouzakis<grbzks@gmail.com> wrote:
On Fri, Jun 26, 2009 at 3:42 AM, Gergely Imreh<imrehg@gmail.com> wrote:
Since i started this, even by stupidly replying to another thread, i might as well answer to that. My suggestion is not having comments in the AUR at all, comments arent useful to the users. They are only useful to the maintainer.
I would disagree... Sometimes it is good for maintaner-user feedback as well. E.g. one of my packages takes a long time to compile. It's a small package but one step looks as if it hung and stays there for about 10-15mins on my computer. I had one of the users place a comment that his compilation didn't work, it froze and he had to kill it after about 5 minutes. Told him to wait a bit longer and it worked. Sure it could be done in the BBS - but would be completely inefficient. Not many users for most of the packages so not many people know what's going on, and I'm not going to search through the forums every day to see if someone wrote about them... Now: comment placed, me notified, can act on it....
Scrap the bbs. If he mailed you, wouldnt you have bothered to reply his email? Or it needs to be done by a comment or not done at all? You wouldnt have added it to the package notes because it wouldnt interest the users? If you had, he wouldnt even have to email you.
Sometimes the maintainer in this case is a Dev or a TU, and eventually they are full of work and they doesn't reply the emails or let the packages out of date or with bugs (in the AUR history *this* fact is happening), what should do the user in this case?: 1.- send an email to the ml? --> will be ignored by everypeople. (and probably the maintainer, and plus force the user to subscribe to the ml, maybe the user isn't a packager, just an user reporting anything like "this package shouldn't depend on bla bla") 2.- "open a bug report", come on dude, not every people will register to the bt to fill a bug report, maybe if one account will work for everything, but this is not the case. 3.- Open a bbs thread (pfff hehee .. I just will laugh about it) 4.- IRC? (why if maintainer isn't active on that?) And plus, sometimes Comments helps the maintainer of a package a LOT. So, this discussion is irrelevant, please when you do an upgrade you don't remove features at least these are unuseful, and reading this thread, more people are agree to keep the comments. Thanks -- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
2009/6/26 Angel Velásquez <angvp@archlinux.com.ve>:
Sometimes the maintainer in this case is a Dev or a TU, and eventually they are full of work and they doesn't reply the emails or let the packages out of date or with bugs (in the AUR history *this* fact is happening), what should do the user in this case?:
1.- send an email to the ml? --> will be ignored by everypeople. (and probably the maintainer, and plus force the user to subscribe to the ml, maybe the user isn't a packager, just an user reporting anything like "this package shouldn't depend on bla bla") 2.- "open a bug report", come on dude, not every people will register to the bt to fill a bug report, maybe if one account will work for everything, but this is not the case. 3.- Open a bbs thread (pfff hehee .. I just will laugh about it) 4.- IRC? (why if maintainer isn't active on that?)
And plus, sometimes Comments helps the maintainer of a package a LOT.
So, this discussion is irrelevant, please when you do an upgrade you don't remove features at least these are unuseful, and reading this thread, more people are agree to keep the comments.
Thanks
I already replied to these many times. Do you know of another project that enables user comments? I know of none. Yet they seem to get by fine with mailing lists, IRC channels, and personal emails. Anyway: http://bugs.archlinux.org/task/15265 -- Greg
On Fri, Jun 26, 2009 at 4:19 AM, Grigorios Bouzakis<grbzks@gmail.com> wrote:
2009/6/26 Angel Velásquez <angvp@archlinux.com.ve>:
Sometimes the maintainer in this case is a Dev or a TU, and eventually they are full of work and they doesn't reply the emails or let the packages out of date or with bugs (in the AUR history *this* fact is happening), what should do the user in this case?:
1.- send an email to the ml? --> will be ignored by everypeople. (and probably the maintainer, and plus force the user to subscribe to the ml, maybe the user isn't a packager, just an user reporting anything like "this package shouldn't depend on bla bla") 2.- "open a bug report", come on dude, not every people will register to the bt to fill a bug report, maybe if one account will work for everything, but this is not the case. 3.- Open a bbs thread (pfff hehee .. I just will laugh about it) 4.- IRC? (why if maintainer isn't active on that?)
And plus, sometimes Comments helps the maintainer of a package a LOT.
So, this discussion is irrelevant, please when you do an upgrade you don't remove features at least these are unuseful, and reading this thread, more people are agree to keep the comments.
Thanks
I already replied to these many times.
Do you know of another project that enables user comments? I know of none. Yet they seem to get by fine with mailing lists, IRC channels, and personal emails.
See for example a project somewhat close to the AUR: http://www.vim.org/scripts/script.php?script_id=2690
Anyway: http://bugs.archlinux.org/task/15265
-- Greg
-- Greg
On Fri, Jun 26, 2009 at 8:49 PM, Grigorios Bouzakis<grbzks@gmail.com> wrote:
2009/6/26 Angel Velásquez <angvp@archlinux.com.ve>:
Sometimes the maintainer in this case is a Dev or a TU, and eventually they are full of work and they doesn't reply the emails or let the packages out of date or with bugs (in the AUR history *this* fact is happening), what should do the user in this case?:
1.- send an email to the ml? --> will be ignored by everypeople. (and probably the maintainer, and plus force the user to subscribe to the ml, maybe the user isn't a packager, just an user reporting anything like "this package shouldn't depend on bla bla") 2.- "open a bug report", come on dude, not every people will register to the bt to fill a bug report, maybe if one account will work for everything, but this is not the case. 3.- Open a bbs thread (pfff hehee .. I just will laugh about it) 4.- IRC? (why if maintainer isn't active on that?)
And plus, sometimes Comments helps the maintainer of a package a LOT.
So, this discussion is irrelevant, please when you do an upgrade you don't remove features at least these are unuseful, and reading this thread, more people are agree to keep the comments.
Thanks
I already replied to these many times.
Do you know of another project that enables user comments? I know of none. Yet they seem to get by fine with mailing lists, IRC channels, and personal emails.
That's why I like Arch, because is different than other projects, if you want those features, then go to "other projects"
I will post there :)
-- Greg
-- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
On Thursday 25 June 2009 21:23:22 Grigorios Bouzakis wrote:
If the user is too lazy to follow the maintainer link and then the link to his email, then he shouldnt be motivated enough to comment either.
Are we trying to package software or testing how motivated our users are? Why make things harder *intentionally*?
If there is need for further discussion, the script is horribly outdated and unmaintained, should be done, like its done now, on the mailing list or the bbs.
How's that for "KISS" ?
Awful? As a small-time maintainer I love that it's easy to tell me when my packages have problems. Reporting a problem is not this bizarre test of motivation: "he is not motivated enough so I don't want his comments", it's just everyday usage! And yes, deleting comments is useful because after the problem is fixed it'just takes space. A *very* simple bug tracker may do, too. -- ("\''/").__..-''"`-. . Roberto Alsina `9_ 9 ) `-. ( ).`-._.`) KDE Developer (MFCH) (_Y_.)' ._ ) `._`. " -.-' http://lateral.netmanagers.com.ar _..`-'_..-_/ /-'_.' The 6,855th most popular site of Slovenia (l)-'' ((i).' ((!.' according to alexa.com (27/5/2007) "Our opponent is an alien starship packed with atomic bombs, I said. We have a protractor. Okay, I’ll go home and see if I can scrounge up a ruler and a piece of string." — Neal Stephenson
I can't believe this conversation is still going! I am fairly certain about two things 1) comments will not be removed 2) comments section may be improved (threaded, ability of maintainer to delete dealt with comments, other?). How did I come up with that conclusion? It seems to be the opinion of one of the (very few) AUR developers who has posted in this thread. And given nothing changes without a patch... Allan
On Fri, Jun 26, 2009 at 4:11 AM, Allan McRae<allan@archlinux.org> wrote:
I can't believe this conversation is still going!
I sent an epiloue 2 mins ago. Check your inbox.
I am fairly certain about two things 1) comments will not be removed 2) comments section may be improved (threaded, ability of maintainer to delete dealt with comments, other?).
How did I come up with that conclusion? It seems to be the opinion of one of the (very few) AUR developers who has posted in this thread. And given nothing changes without a patch...
Allan
-- Greg
2) comments section may be improved (threaded, ability of maintainer to delete dealt with comments, other?).
Code blocks. Where's the code for the AUR? Maybe I can submit some patches.
On Fri, Jun 26, 2009 at 10:30 AM, Xyne<xyne@archlinux.ca> wrote:
2) comments section may be improved (threaded, ability of maintainer to delete dealt with comments, other?).
Code blocks.
Where's the code for the AUR? Maybe I can submit some patches.
http://projects.archlinux.org/?p=aur.git;a=summary You should also subscribe to the aur-dev ML -- Callan Barrett
If the interface were to provide notification when and if the maintainer has linked a url under, say, a "discussion thread" variable, then it'll work out pretty much the same way as how comments are handled now. But this also means the maintainer has to know whether someone actually opened up a thread. In the end, both ways are KISS as one proposes a separate location for discussion-related content. I for one, have no problem whether comments remain or not. From a user-experience perspective, it is definitely worth it. Looking at it practically, since it's an almost-public database, comments serve more purposes than just information or peer-review. But when the informational posts escalate into a prolonged discussion, is where the problem comes in. Also, I suppose the notification for comments is currently used as a workaround for (the lack of) notification on updates for those not using a local front-end.
On Thu, 25 Jun 2009 17:50:34 +0800 Ray Rashif <schivmeister@gmail.com> wrote:
If the interface were to provide notification when and if the maintainer has linked a url under, say, a "discussion thread" variable, then it'll work out pretty much the same way as how comments are handled now. But this also means the maintainer has to know whether someone actually opened up a thread. In the end, both ways are KISS as one proposes a separate location for discussion-related content.
I for one, have no problem whether comments remain or not. From a user-experience perspective, it is definitely worth it. Looking at it practically, since it's an almost-public database, comments serve more purposes than just information or peer-review. But when the informational posts escalate into a prolonged discussion, is where the problem comes in.
Also, I suppose the notification for comments is currently used as a workaround for (the lack of) notification on updates for those not using a local front-end.
I don't really see a point in that much discussion. [community] will vanish from AUR anyway. 1) [unsupported] should keep the comments. 2) The maintainer of a package should be able to delete comments (cleanup possibility). 3) Notification should be enabled by default (or add a switch for the default behaviour). Regards, Philipp
2009/6/25 <hollunder@gmx.at>:
On Thu, 25 Jun 2009 17:50:34 +0800 Ray Rashif <schivmeister@gmail.com> wrote:
I don't really see a point in that much discussion. [community] will vanish from AUR anyway.
1) [unsupported] should keep the comments.
2) The maintainer of a package should be able to delete comments (cleanup possibility).
3) Notification should be enabled by default (or add a switch for the default behaviour).
Regards, Philipp
I agree. Personally, I find filing a bug quite daunting (I know it isn't but I do) and it is easier for me, as the bug reporter, to just shove my problem in the box at the bottom. As the maintainer, as most of the bugs I get reported are simple, and I think that's the case with the vast majority of AUR packages, it is easier to just have a simple interface little comment interface, instead of a massive, full blown bug management system. The maintainer of a package should be able to delete comments as this would make some of the more popular packages' comments easier to clean up. Instead of a trusted user needing to do this, the owner of the package could be allowed to delete any comments older that a week. That's a policy decision, but I hope we will have some way of deleting comments that refer to bugs that have been fixed. Notification should be enabled as default, as I find with some packages people report them out of date if there is any kind of problem with them. This leads to confusion, as to whether the package is actual out of date, or if it is just a problem with the PKGBUILD. Maybe a bug button for people writing comments would be useful, so the comment could trigger an email alert, whereas a normal comment might not. All of this, in my opinion would still be easier than a bug tracker. Thanks, Laurie
Yes, and these should be filed as feature requests for AUR ;)
Not much point filing them as feture requests if you (as in the collective, not you in person) decide to deleat the comment system xD. The point was that reform is better than change (oooh I sound so concervative). Do what you want xD, Laurie 2009/6/25 Ray Rashif <schivmeister@gmail.com>:
Yes, and these should be filed as feature requests for AUR ;)
The maintainer of a package should be able to delete comments as this would make some of the more popular packages' comments easier to clean up. Instead of a trusted user needing to do this, the owner of the package could be allowed to delete any comments older that a week. That's a policy decision, but I hope we will have some way of deleting comments that refer to bugs that have been fixed.
This would prevent users from being able to flag malicious packages. If this were implemented, I would like to see a "report malicious package" link or something else. As Arch continues to grow we will end up with malicious users and I would prefer to be prepared to handle these when the time comes.
2009/6/25 Xyne <xyne@archlinux.ca>:
The maintainer of a package should be able to delete comments as this would make some of the more popular packages' comments easier to clean up. Instead of a trusted user needing to do this, the owner of the package could be allowed to delete any comments older that a week. That's a policy decision, but I hope we will have some way of deleting comments that refer to bugs that have been fixed.
This would prevent users from being able to flag malicious packages. If this were implemented, I would like to see a "report malicious package" link or something else. As Arch continues to grow we will end up with malicious users and I would prefer to be prepared to handle these when the time comes.
That's why I said "delete any comments older that a week". From what I have seen, almost all packages are checked, and I can't imagine that anyone who found a malicious package wouldn't report it, if not here, to the forums. In either case, it would be discovered, as I doubt that if the issue has not been brought to the attention of the community in the week after discovery then I doubt it will be in any sensible timescale.
On Thu, 25 Jun 2009 19:27:43 +0000 Laurie Clark-Michalek <bluepeppers@archlinux.us> wrote:
2009/6/25 Xyne <xyne@archlinux.ca>:
The maintainer of a package should be able to delete comments as this would make some of the more popular packages' comments easier to clean up. Instead of a trusted user needing to do this, the owner of the package could be allowed to delete any comments older that a week. That's a policy decision, but I hope we will have some way of deleting comments that refer to bugs that have been fixed.
This would prevent users from being able to flag malicious packages. If this were implemented, I would like to see a "report malicious package" link or something else. As Arch continues to grow we will end up with malicious users and I would prefer to be prepared to handle these when the time comes.
That's why I said "delete any comments older that a week". From what I have seen, almost all packages are checked, and I can't imagine that anyone who found a malicious package wouldn't report it, if not here, to the forums. In either case, it would be discovered, as I doubt that if the issue has not been brought to the attention of the community in the week after discovery then I doubt it will be in any sensible timescale.
Sorry, I missed the "older than a week" part. That should be more than enough time for the package to have been reported and deleted. Ignore my previous reply.
Am Fri, 26 Jun 2009 00:52:46 +0200 schrieb Xyne <xyne@archlinux.ca>:
Sorry, I missed the "older than a week" part. That should be more than enough time for the package to have been reported and deleted. Ignore my previous reply.
One week is not enough. It must be at least one month. If a maintainer is on vacation, it's not unusual, that he can't read comments for his package for some weeks. And there has to be a button to prevent important comments from automatical deletion. But I still prefer the delete button for the maintainers, so that comments can be deleted manually by the maintainers. Cheers, Heiko
Sorry, I missed the "older than a week" part. That should be more than enough time for the package to have been reported and deleted. Ignore my previous reply.
One week is not enough. It must be at least one month. If a maintainer is on vacation, it's not unusual, that he can't read comments for his package for some weeks.
And there has to be a button to prevent important comments from automatical deletion.
But I still prefer the delete button for the maintainers, so that comments can be deleted manually by the maintainers.
Cheers, Heiko
I think you misunderstood my reply. It was not about comments being automatically pruned after a week. It was referring to a maintainer's ability to delete comments. If the maintainer is there to delete them, he is also there to read them. The minimum of one week before a comment can be deleted would prevent the following situation: Alice detects that Eve's package is malicious. Alice leaves a comment on the AUR warning others. Alice contacts AUR-general to get a TU to delete the package. Eve deletes the warning. Bob installs the malicious package because he didn't see the warning. A TU sees Alice's message a few hours later and deletes Eve's package. A TU will certainly get Alice's message within a week of posting it so the Eve's package will get deleted before Eve can remove Alice's warning. Bob and others will therefore avoid installing the package before it has been deleted.
Am Fri, 26 Jun 2009 04:45:16 +0200 schrieb Xyne <xyne@archlinux.ca>:
I think you misunderstood my reply. It was not about comments being automatically pruned after a week. It was referring to a maintainer's ability to delete comments. If the maintainer is there to delete them, he is also there to read them.
Well, I see. I think I have indeed misunderstood your reply. ;-) I'm not sure, if it's really necessary, but could make sense. Maybe such a deletion blocker could only be implemented for malicious reports. See the "report malicious code" button idea, which would post a comment to the package and send an e-mail to this mailing list and/or the TUs directly. Those "comments" could be blocked for at least a week. And alternative could be, that only TUs get a button to remove the warning after the maintainer has fixed the package and a TU has reviewed the new code. Then such a timed blocker wasn't necessary. And maybe the downloadlinks should be blocked at least for users (not for TUs and the maintainer), if such a malicious code warning was sent, if this is possible to implement. It's a bit like the prior "package is save" function, but limited to packages, which are reported as malicious. I hope there won't be too many of them. Cheers, Heiko
On Fri, 26 Jun 2009 00:52:46 +0200 Xyne <xyne@archlinux.ca> wrote:
On Thu, 25 Jun 2009 19:27:43 +0000 Laurie Clark-Michalek <bluepeppers@archlinux.us> wrote:
2009/6/25 Xyne <xyne@archlinux.ca>:
The maintainer of a package should be able to delete comments as this would make some of the more popular packages' comments easier to clean up. Instead of a trusted user needing to do this, the owner of the package could be allowed to delete any comments older that a week. That's a policy decision, but I hope we will have some way of deleting comments that refer to bugs that have been fixed.
This would prevent users from being able to flag malicious packages. If this were implemented, I would like to see a "report malicious package" link or something else. As Arch continues to grow we will end up with malicious users and I would prefer to be prepared to handle these when the time comes.
That's why I said "delete any comments older that a week". From what I have seen, almost all packages are checked, and I can't imagine that anyone who found a malicious package wouldn't report it, if not here, to the forums. In either case, it would be discovered, as I doubt that if the issue has not been brought to the attention of the community in the week after discovery then I doubt it will be in any sensible timescale.
Sorry, I missed the "older than a week" part. That should be more than enough time for the package to have been reported and deleted. Ignore my previous reply.
IMHO malicious packages should be reported to the list anyway to be removed ASAP. A comment alone wouldn't do it anyway. Regards, Philipp
Am Fri, 26 Jun 2009 02:57:23 +0200 schrieb <hollunder@gmx.at>:
On Fri, 26 Jun 2009 00:52:46 +0200 Xyne <xyne@archlinux.ca> wrote:
IMHO malicious packages should be reported to the list anyway to be removed ASAP. A comment alone wouldn't do it anyway.
Principally you are right, but pressing a button "report malicious package" could or should send an e-mail to this mailing list or to every TU automatically. This would be the easiest way for the users. Cheers, Heiko
On Fri, Jun 26, 2009 at 8:59 PM, Heiko Baums<lists@baums-on-web.de> wrote:
Am Fri, 26 Jun 2009 02:57:23 +0200 schrieb <hollunder@gmx.at>:
On Fri, 26 Jun 2009 00:52:46 +0200 Xyne <xyne@archlinux.ca> wrote:
IMHO malicious packages should be reported to the list anyway to be removed ASAP. A comment alone wouldn't do it anyway.
Principally you are right, but pressing a button "report malicious package" could or should send an e-mail to this mailing list or to every TU automatically. This would be the easiest way for the users.
This will save the time to write a long e-mail to the ml writting "the package X which are in http://xxx is malicious" but this feature should be powerful if you add a "why" this package were reported as malicious, because doing accidentally clicks can send an inoportunous e-mail to the ml. -- Angel Velásquez angvp @ irc.freenode.net Linux Counter: #359909
Am Fri, 26 Jun 2009 21:03:06 +1930 schrieb Angel Velásquez <angvp@archlinux.com.ve>:
This will save the time to write a long e-mail to the ml writting "the package X which are in http://xxx is malicious" but this feature should be powerful if you add a "why" this package were reported as malicious, because doing accidentally clicks can send an inoportunous e-mail to the ml.
This can easily be done and solved by loading a second page with a text field for the "why" and a "confirm" button similar to the "flag as out of date" function in the package search (http://www.archlinux.org/packages/). The name and e-mail address from the reporter should be added to the e-mail. Cheers, Heiko
Grigorios, you're an idiot. You're basically just trolling the thread now and, correct me if I'm wrong, have provided nothing in the way of patches for the AUR like open source is magic where you whine about stuff and it gets fixed how you like. I see absolutely no compelling reason to get rid of comments and it's ridiculous this thread even exists. The only interesting thing is the idea of improving the comment system and there are no patches for that either. Grigorios: stop digging yourself into a hole. -- Callan Barrett
On Fri, Jun 26, 2009 at 4:36 AM, Callan Barrett<wizzomafizzo@gmail.com> wrote:
Grigorios, you're an idiot. You're basically just trolling the thread now and, correct me if I'm wrong, have provided nothing in the way of patches for the AUR like open source is magic where you whine about stuff and it gets fixed how you like.
I already said that i cant provide patches for the AUR. And i said that it was a suggestiong. I also already said i wasnt willing to continue talking about my suggestion on the mailing list anymore since people dont like it. After that, people who hadnt yet taken part in the discussion appeared, you and Angel for example.
I see absolutely no compelling reason to get rid of comments and it's ridiculous this thread even exists. The only interesting thing is the idea of improving the comment system and there are no patches for that either.
So now you admit something isnt working as it should and should/could improve. And then you are calling me an idiot for making a suggestion on how to improve that, before anyone else mentioned it. Others didnt like my suggestion. What more do you want? Apologise for making it? Start a new discussion thread if you like, about *altering* the way comments work in the AUR since thats what you think is more appropriate. Close the feature request as Wont implement since you dont think its right.
Grigorios: stop digging yourself into a hole.
-- Callan Barrett
-- Greg
On Fri, Jun 26, 2009 at 9:57 AM, Grigorios Bouzakis<grbzks@gmail.com> wrote:
I see absolutely no compelling reason to get rid of comments and it's ridiculous this thread even exists. The only interesting thing is the idea of improving the comment system and there are no patches for that either.
So now you admit something isnt working as it should and should/could improve. And then you are calling me an idiot for making a suggestion on how to improve that, before anyone else mentioned it. Others didnt like my suggestion. What more do you want? Apologise for making it? Start a new discussion thread if you like, about *altering* the way comments work in the AUR since thats what you think is more appropriate. Close the feature request as Wont implement since you dont think its right.
Yes, when things could use improvement the only good way to deal with it is to remove it outright. You're the fountain of wisdom. I don't know what you should do. Start your own AUR fork called funAUR? Start a blog where you list features in applications that should be removed for not being perfect? Write a semi-successful linux-based webcomic where you're the angsty main character? As long as you stop posting on the mailing list. I will gladly close your bug. -- Callan Barrett
Am Fri, 26 Jun 2009 10:08:44 +0800 schrieb Callan Barrett <wizzomafizzo@gmail.com>:
I will gladly close your bug.
Loui has already done this. :-) Cheers, Heiko
Grigorios, you're an idiot. You're basically just trolling the thread now and, correct me if I'm wrong, have provided nothing in the way of patches for the AUR like open source is magic where you whine about stuff and it gets fixed how you like.
I see absolutely no compelling reason to get rid of comments and it's ridiculous this thread even exists. The only interesting thing is the idea of improving the comment system and there are no patches for that either.
Grigorios: stop digging yourself into a hole.
Although I strongly disagree with Greg and think that he is pursuing this in an obnoxious manner, I think your reply is a bit over the top. There's no need to call people idiots here. Also, and this is directed at devs in general, the whole "stfu unless you can submit a patch" mentality is narrow-minded at best. Just because someone can't code it themselves doesn't mean that an idea is bad.Too many people act as though a suggestion were a demand and many devs content themselves with "good enough" if they can avoid putting more work into it. Not everyone is a coder and it would often take someone far more time and effort to create a patch than a dev who's intimately familiar with the code. It's the same thing as humming a tune to a musician for a bridge in his song. If you don't play any instruments and he tells you to come back with a recording of it, nobody benefits. There's some quote about how the "good enough" mentality is the greatest barrier to progress. If someone knows it, please post it.
Xyne wrote:
Also, and this is directed at devs in general, the whole "stfu unless you can submit a patch" mentality is narrow-minded at best. Just because someone can't code it themselves doesn't mean that an idea is bad.Too many people act as though a suggestion were a demand and many devs content themselves with "good enough" if they can avoid putting more work into it. Not everyone is a coder and it would often take someone far more time and effort to create a patch than a dev who's intimately familiar with the code.
I think you will notice that that attitude tends to appear once a thread has reached 50 comments... It is generally intended to be a "stfu" rather than a "stfu unless you can submit a patch". This will be the 75th post in the thread and I don't think a new point has been made since about post three. Allan
Xyne wrote:
Also, and this is directed at devs in general, the whole "stfu unless you can submit a patch" mentality is narrow-minded at best. Just because someone can't code it themselves doesn't mean that an idea is bad.Too many people act as though a suggestion were a demand and many devs content themselves with "good enough" if they can avoid putting more work into it. Not everyone is a coder and it would often take someone far more time and effort to create a patch than a dev who's intimately familiar with the code.
I think you will notice that that attitude tends to appear once a thread has reached 50 comments... It is generally intended to be a "stfu" rather than a "stfu unless you can submit a patch". This will be the 75th post in the thread and I don't think a new point has been made since about post three.
Allan
There are a few bits and pieces that have come out of it but yeah, overall it should be TGN'd. Still, a better reply would have been "you're alone on this, let it go" instead of "so submit a patch". Meh. Xyne
Principally you are right, but pressing a button "report malicious package" could or should send an e-mail to this mailing list or to every TU automatically. This would be the easiest way for the users.
That could lead to spam. A better system would be similar to the out-of-date system that we currently have, with some changes. You press the "report malicious package" button, submit a reason, and then a messages gets automatically posted to the list. At the same time, it also displays on the AUR page and flagged packages can be filtered in the search the same way out-of-date packages can. The reporter would also be mentioned in the list (to prevent people from anonymously flagging packages without reason).
On Thu, Jun 25, 2009 at 10:02 PM, Xyne<xyne@archlinux.ca> wrote:
The maintainer of a package should be able to delete comments as this would make some of the more popular packages' comments easier to clean up. Instead of a trusted user needing to do this, the owner of the package could be allowed to delete any comments older that a week. That's a policy decision, but I hope we will have some way of deleting comments that refer to bugs that have been fixed.
This would prevent users from being able to flag malicious packages. If this were implemented, I would like to see a "report malicious package" link or something else. As Arch continues to grow we will end up with malicious users and I would prefer to be prepared to handle these when the time comes.
You mean you have seen comments about malicious packages and removed them? I doubt it. Malicious or useless packages, at least since mere users are not able to delete them, have *always* been reported on the mailing list or to TUs directly, or the AUR cleanup wiki page. -- Greg
participants (19)
-
Aaron Griffin
-
Alexander Rødseth
-
Allan McRae
-
Andrei Thorp
-
Angel Velásquez
-
bardo
-
Callan Barrett
-
Daenyth Blank
-
Gergely Imreh
-
Grigorios Bouzakis
-
Heiko Baums
-
hollunder＠gmx.at
-
Laurie Clark-Michalek
-
Loui Chang
-
Ray Rashif
-
Roberto Alsina
-
Ronald van Haren
-
Xavier
-
Xyne